Ultrastructural changes of sparkling wine lees during long-term aging in real enological conditions


  • Rebeca Tudela,

    1. Departament de Nutrició i Bromatologia, Xarxa de Referència en Tecnologia dels Aliments (XaRTA), Facultat de Farmàcia, Universitat de Barcelona, Barcelona, Spain
    Search for more papers by this author
  • Joan J. Gallardo-Chacón,

    1. Institut Universitaire de la Vigne et du Vin ‘Jules Guyot’, EA 581 EMMA, Université de Bourgogne, Dijon, France
    Search for more papers by this author
  • Núria Rius,

    Corresponding author
    1. Departament de Microbiologia i Parasitologia Sanitàries, Facultat de Farmàcia, Universitat de Barcelona, Barcelona, Spain
    • Departament de Nutrició i Bromatologia, Xarxa de Referència en Tecnologia dels Aliments (XaRTA), Facultat de Farmàcia, Universitat de Barcelona, Barcelona, Spain
    Search for more papers by this author
  • Elvira López-Tamames,

    1. Departament de Nutrició i Bromatologia, Xarxa de Referència en Tecnologia dels Aliments (XaRTA), Facultat de Farmàcia, Universitat de Barcelona, Barcelona, Spain
    Search for more papers by this author
  • Susana Buxaderas

    1. Departament de Nutrició i Bromatologia, Xarxa de Referència en Tecnologia dels Aliments (XaRTA), Facultat de Farmàcia, Universitat de Barcelona, Barcelona, Spain
    Search for more papers by this author

Correspondence: Núria Rius, Departament de Microbiologia i Parasitologia Sanitàries, Facultat de Farmàcia, Universitat de Barcelona, Avda. Joan XXIII s/n, 08028 Barcelona, Spain. Tel.: +34 93 402 4497; fax: +34 93 402 4498; e-mail: nrius@ub.edu


Ultrastructural changes of lees of three series of sparkling wines produced using the traditional method during long-term aging (4 years) were assessed by high-pressure freezing in combination with transmission electron microscopy. The stratified structure of the cell wall disappeared throughout aging. After 18 months, the microfibrous material of the cell wall appeared more diffuse and the amorphous midzone of the inner wall layer was progressively degraded. From 30 months onward, the cell wall consisted of a tangled structure of fibers. In spite of these changes, the cell wall of yeasts remained unbroken at 48 months of wine aging. Cell membrane breakage was observed for the first time in lees of Saccharomyces cerevisiae. An increase in the thickness of the periplasmic space owing to plasmolysis and of the number of cells with less cytoplasmic content was observed during aging. Morphological evidence of microautophagy was detected for the first time in S. cerevisiae in enological conditions.


Sparkling wines are distinguished from other wines because they undergo two fermentation processes: grapes must ferment to produce wine and then, after the addition of the tirage liquor, the wine itself is fermented. They also undergo biological aging which consists of leaving the lees of the second fermentation in contact with the sparkling wine in the same bottle or tank where the fermentation is carried out (Official Journal of the European Union L 148/47). Biological aging duration varies, but for quality sparkling wines of European origin (e.g. Champagne in France, Cava in Spain, Sekt in Germany, and Franciacorta, Asti, and Prosecco in Italy) it usually takes anywhere from a few months to several years. In the case of Cava, the minimum aging period is 9 months, with 15 months for the Reserve and 30 months for the Great Reserve. Biological aging gives to these wines particular organoleptic characteristics owing to their contact with the lees. It is assumed that the sensorial profile acquired by these wines is because of the autolytic process of cells during aging (Torrens et al., 2010). Many researchers have tried to connect the changes observed in the sparkling wine during aging with autolysis by analyzing the evolution of many compounds in the wine over time. When an increase was observed in the content of cell compounds, such as proteins, peptides, amino acids, fatty acids, polysaccharides, etc., during aging, it was suggested that the cells had released these compounds into the wine as a result of cellular self-destruction (Feuillat & Charpentier, 1982; Moreno-Arribas et al., 1996; Fornairon-Bonnefond et al., 2002; Gallart et al., 2002; Charpentier et al., 2005; Pérez-Serradilla & Luque de Castro, 2008; Martínez-Rodriguez & Pueyo, 2009). Other compositional changes were postulated as a result of the activity of enzymes released by the dead cells (Leroy et al., 1990; Alexandre et al., 2001).

Although induced autolysis is widely used (Fornairon-Bonnefond et al., 2002; Zhao & Fleet, 2003, 2005; Alexandre & Guilloux-Benatier, 2006), few studies focusing on autolysis have been made during the evolution of sparkling wine fermented and aged in the bottle, and under real biological aging conditions, temperature often below 15 °C and under a carbon dioxide pressure of 6 atmospheres (Piton et al., 1988; Leroy et al., 1990; Martínez-Rodríguez et al., 2001). These biological aging conditions of sparkling wines are far from the optimal conditions of yeast autolysis (pH 5.0, and temperature 45 °C) (Charpentier, 2010).

In the last 20 years, enological additives have been developed based on inactive dry yeast with technological and sensorial purposes (Pozo-Bayón et al., 2009a, 2009b). The development of these products has advanced the knowledge of the structural composition of the wall and other cellular organelles, as well as many of the structural changes suffered by the dead yeast cells when autolysis was induced. However, most of these studies were carried out in model wines where lees were subjected to accelerated autolysis. Such induced autolysis was not comparable with the natural autolysis that occurs during the biological aging of sparkling wines, as has been reported by Martínez-Rodríguez et al. (2001).

There is a single published study on the structural changes of the second fermentation lees of a bottle-fermented sparkling wine during long-term aging. Piton et al. (1988) used thin-section electron microscopy to observe the structural changes of lees from some champagne bottles with different aging. These authors observed that most aged lees have a thinner wall than younger dead cells and that even in one 7-year-old sample the cell wall of the lees remained unbroken.

The aim of this study was to demonstrate the evolution of structural changes and cellular degradation of lees during long-term aging in real enological conditions. For this purpose, high-pressure freezing (HPF) in combination with transmission electron microscopy (TEM) was used to monitoring the ultrastructural changes of lees of three series of sparkling wines produced with the same Saccharomyces cerevisiae strain and aged for 4 years in bottles.

Materials and methods


Three sparkling wines series from three different blends were produced on an industrial scale at Freixenet S.A. winery using the traditional method. The three base wines were bottled after the addition of the same ‘liqueur de tirage’. This liqueur is composed of 1–2 × 106 yeast cells mL−1 (yeast starter), 500 g L−1 of sucrose, 0.1–0.2 g L−1 of bentonite, and ammonium phosphate (50 mL hL−1). Taking into account that approximately 40–42 mL of ‘liqueur de tirage’/750 mL bottle is added, each bottle receives between 20 and 21 g of sucrose (Flanzy et al., 1999). The S. cerevisiae F11 strain belongs to the private collection of Freixenet S.A. winery. Cava series (500 bottles of 750 mL × 3 series) were stored in the cellar for 48 months during the second fermentation ‘prise de mousse' and aging. Three bottles of each series were taken at 1, 9, 13, 18, 21, 30, and 48 months after bottling. These sampling points were selected according to representative aging periods of Cava categories: (1) Cava (≥ 9 months), (2) Reserve (≥ 15 months), and (3) Great Reserve (≥ 30 months). In addition, a yeast starter sample for the second fermentation was collected as a reference to recognize the ultrastructures of living cells (time 0).

Isolation of lees

The lees were prepared as follows: the content of three bottles (750 mL each) was centrifuged at 8350 g for 15 min at 4 °C using a Centrikon T-124 centrifuge (Kontron Instruments, Cumbernauld, UK); the pellet was washed three times with 30 mL of NaCl 0.9% (Panreac, Barcelona, Spain) (Leroy et al., 1990).

Transmission electron microscopy

Lees pellets were selected under a stereomicroscope and transferred to 1.5 mm diameter and 200-μm deep planchettes and immediately cryoimmobilized using a Leica EMPact high-pressure freezer (Leica, Vienna, Austria), and then stored in liquid nitrogen until further use (Nevot et al., 2006). After rapid freezing, specimens were freeze-substituted in anhydrous acetone containing 2% osmium tetroxide and 0.1% uranyl acetate at −90 °C for 3 days. They were gradually warmed to room temperature at a temperature progression rate of 5 °C h−1. After several acetone rinses, the samples were embedded in epon resin. The epon was then polymerized for 48 h at 60 °C (Walther & Ziegler, 2002). Ultrathin sections were cut using a Leica Ultracut UCT ultra microtome, mounted on formvar-coated copper grids and poststained with 2% uranyl acetate in water and lead citrate (Bozzola & Russell, 1999). The ultrathin sections were observed with a Tecnai Spirit electron microscope (FEI Co, Hillsboro, OR) at an accelerating voltage of 120 kV. On average, 10 overview and approximately 55–65 detailed electron micrographs for each sample point were taken.

Identification of yeast organelles and structure has been carried out on the basis of the TEM recognition studies on living cells reported by Wright (2000), Walther & Ziegler (2002), Giddings (2003), Osumi et al. (2006), and Yamaguchi et al. (2011).

Results and discussion

HPF in combination with TEM was used to study the ultrastructural changes in yeast cells along with sparkling wine aging on lees. Cryofixation by HPF preserves cell ultrastructure close to the normal living state because the intrinsic cell water turns into vitreous ice and the cell content is rapidly and efficiently immobilized (Studer et al., 2008).

No differences were observed among the lees of the three series of cava (data not shown), although the base wines were different.

The ultrastructure of the S. cerevisiae strain used in this study is shown in Fig. 1(time 0). The cell wall consisted of an amorphous electron-transparent inner layer (iCW) and an electron-dense outer layer (oCW) (Fig. 1a and d) (Walther & Ziegler, 2002; Yamaguchi et al., 2011). Branched β1,3-glucan chains containing β1,6-linked glucose residues could be part of the amorphous layer (AL). The outer layer is observed as an accumulated microfibrous material that extends into the medium. It consists predominantly of mannoproteins. Enzymes and other proteins are physically entrapped in the wall (Kopecká et al., 1974; Cid et al., 1995; Lipke & Ovalle, 1998; Klis et al., 2002, 2006; Yamaguchi et al., 2011). Figure 1b and c reveal the nucleus (N), nucleolus (Nu), nuclear pore (NP), spindle pole body (SPB), and the nuclear envelope (NE) (Wright, 2000; Walther & Ziegler, 2002; Osumi et al., 2006; Yamaguchi et al., 2011). Ribosomes (R) (Fig. 1a) (Osumi et al., 2006; Yamaguchi et al., 2011), a large vacuole (V) with polyphosphate aggregates (Po) (Fig. 1b), glycogen granules (GLY) (Fig. 1c) (Osumi, 1998), and mitochondria (M), cytoplasmic membrane (CM) and invagination of cell membrane (CMI) (Fig. 1a and d) can also be observed (Walther & Ziegler, 2002; Giddings, 2003; Osumi et al., 2006; Yamaguchi et al., 2011).

Figure 1.

TEM images of Saccharomyces cerevisiae cells cryofixed by HPF. Note that cell structures were well preserved in a natural state. Bars: (a) 200 nm, (b) 1 μm, (c) 500 nm, and (d) 200 nm. CM, cell membrane; CMI, cell membrane invaginations; CW, cell wall; iCW, inner cell wall; oCW, outer cell wall; cyt, cytosol; rER, rough endoplasmic reticulum; G, Golgi apparatus; Gly, glycogen granule; L, lipid droplet; M, mitochondrion; N, nucleus; NE, nuclear envelope; NP, nuclear porus; Nu, nucleolus; P, peroxisome, Po, polyphosphate; PS, periplasmic space; R, ribosome; SPB, spindle pole body; V, vacuole; ve, vesicle.

Ultrastructure of yeasts during second fermentation

Figure 2 shows ultrathin sections of S. cerevisiae harvested after 1 month of fermentation in the bottle. Yeast cells at this stage are similar in appearance to yeast starter cells (Fig. 1). No signs of autolysis were observed. The cell wall maintained its two-layer structure attached to the CM, and both the NE and the CM were visible (Fig. 2a, c and d) (Wright, 2000; Walther & Ziegler, 2002; Osumi et al., 2006; Yamaguchi et al., 2011). Many cytoplasmic organelles can be discerned; ribosomes (R) (Fig. 2a, c and d) (Osumi et al., 2006; Yamaguchi et al., 2011); mitochondria (M) with very few cristae, typical of yeasts in anaerobiosis (Piton et al., 1988), and the bilayer structure of the membrane comprising yeast Golgi cisternae (G) (Fig. 2b); a lipid body (L) (Fig. 2a); a peroxisome (P) (Fig. 2c); and many endoplasmic reticulum structures with many ribosomes (rER) (Fig. 2d) (Giddings, 2003; Osumi et al., 2006; Yamaguchi et al., 2011).

Figure 2.

Ultrathin sections at high magnification of Saccharomyces cerevisiae lees after 1 month of bottling. Bilayered structure of cytoplasmic (CM), nuclear envelope (NE), mitochondrion (M), and Golgi (G) membranes is well distinguished. Bars: (a and b) 500 nm, (c and d) 200 nm.

Autophagy of yeasts

Evidence of microautophagy and macroautophagy was detected in sparkling wine lees (Fig. 3). In cells harvested at 1 month of bottling, we were able to observe a large vacuole (V) that formed arm-like extensions of its membrane engulfing a vesicle (ves) (Fig. 3a) and lipid bodies (L) (Fig. 3c). Fusion and concomitant sealing of this growing membrane will result in the incorporation of the vesicle and the lipid body into the vacuole. This microautophagic process has been described for peroxisome degradation in Pichia pastoris (Mijaljica et al., 2007; Kraft et al., 2009). Figure 3a also showed a nonselective autophagy process by which bulk cytosol (cyt) was taken up by tubular invaginations of the vacuole. Lipid bodies (L) were also assimilated by arm-like protrusions (Fig. 3b, arrows) of the vacuole surface. A similar microautophagic process was postulated by Kraft et al. (2009) to take up large cytosol components in starved S. cerevisiae cells in laboratory conditions. Figure 3c shows a yeast cell whose vacuole has formed an autophagic tube (AT). Vacuolar invaginations in nitrogen-starved S. cerevisiae cells were visualized with the viability dye FM 4-64 under a confocal laser scanning microscope by Mijaljica et al. (2007). These authors observed that membrane vacuolar invaginations formed AT that disappeared within one hour of labeling with FM 4-64, leaving a diffuse stain in the vacuole. They suggested that an ‘inverse-budding’ reaction of the AT maybe related to the regulation of vacuolar membrane homeostasis. We propose that the vacuole of the cell shown in Fig. 3c was undergoing a similar microautophagic process to control its own homeostasis. To our knowledge we showed here, for the first time, morphological evidence of industrial S. cerevisiae undergoing microautophagy during real sparkling wine aging.

Figure 3.

Micrographs of lees showing different types of autophagy. (a–d) cells harvested after 1 month of fermentation. (a and b) Microautophagy: the sequestration event occurs at the surface of the vacuole (see text for explanation); (c) vacuole self-degradation by forming an AT; (d) the yeast vacuole contains many autophagic bodies (white arrows), an autophagosome (black arrow) is observed in the cytoplasm of the cell; and (e–g) dead cells harvested at 13 months (e), 21 months (f), and 30 months (g) of aging with an autophagosome (arrows). Bars: (a) 500 nm, (b) 1 μm, (c) 2 μm, (d) 1 μm, (e) 500 nm, (f) 200 nm, (g) 500 nm.

An autophagosome near the vacuole (Fig. 3d, black arrow) and autophagic bodies were observed inside the vacuoles of S. cerevisiae cells (Fig. 3d, white arrows), confirming the findings of Cebollero & Gonzalez (2006) that macroautophagy was also induced during the second fermentation of wine in a closed bottle. Samples taken after 1 month of bottling were filled with living cells that were undergoing autophagy but not yet autolysis. Our results agree well with those reported by Cebollero & Gonzalez (2006) that autophagy takes place before autolysis. The detection of a single and reminiscent autophagosome in the cytoplasm of cells at 13, 21, and 30 months of aging (Fig. 3e, f and g, respectively) could be owing to the resistance of bilayered membranes to degradation (see 'Evolution of cell cytoplasm lees').

After 1 month of bottled fermentation, S. cerevisiae had reached the stationary phase (data provided by Freixenet S.A.), and the growth arrest is the consequence of depletion of nutrients and high levels of ethanol (Bauer & Pretorius, 2000). At this stage, different autophagy processes might be induced to control cellular homeostasis and to contribute to the outcome of autolysis in enological conditions.

Autolysis of yeasts

Lees samples harvested at 9, 18, 30, and 48 months after bottling are shown in Fig. 4a, b, c and d, respectively. From the ninth month onward, most of the cells were plasmolysed; a double CM completely detached from the cell wall was observed (Fig. 4). Although samples were collected after 9 months of bottling to make sure that cells were undergoing autolysis (Charpentier & Feuillat, 1993; Todd et al., 2000), we detected some cells that apparently had not yet begun. Figure 4a insert shows two yeast cells of the same preparation taken at 9 months. One of them (left) is not yet plasmolysed. Yeast autolysis is a slow process associated with cell death, and involves intracellular hydrolytic enzymes that act to release cytoplasmic and cell wall compounds into the wine (Babayan & Bezrukov, 1985). The low temperature at which the aging process in the bottle takes place causes a low death rate and low enzymatic reaction rates, explaining the slowness of the yeast autolysis. Changes in the cell wall were more homogeneous within cells of the same age than the degradation of cytoplasmic material, as can be seen in Fig. 4.

Figure 4.

Low magnification view of ultrathin sections of sparkling wine lees harvested at 9 months (a), 18 months (b), 30 months (c), and 48 months (d) of aging. (a, insert) a completely intact cell was detected after 9 months of aging. Bars: (a–d) 5 μm, (a, inside) 2 μm.

Evolution of the cell walls

Figure 5 shows micrographs (200 nm of magnification) of thin sections of yeasts cells after 1, 9, 13, 18, 30, and 48 months of aging. These micrographs confirmed that the cell wall of the yeasts remained unbroken throughout autolysis. As stated above ('Ultrastructure of yeasts during second fermentation') after 1 month of bottling, the cell wall maintained its two-layer structure consisting of an amorphous inner layer (iCW) and a fibrous outer layer (oCW) (Fig. 5a). At 9 months, the complete detachment of the CM accompanying the plasmolysis of the cells enabled the identification of the stratified structure of the S. cerevisiae cell wall (Fig. 5b). The HPF technique used in this study improved the quality of the micrographs enabling us to observe that the inner layer of the cell wall (iCW) also consisted of two different layers: a microfibrous layer (FL) and an AL. The fibrous layer (FL) was formed of a homogeneous structure of very tight electron-dense microfibers that extended to the periplasmic space (PS). The midzone of the cell wall was observed as an amorphous matrix (AL). The outer cell wall layer (oCW) in cells of samples taken at 9 and 13 months of bottling presented the same microfibrous structure as the yeast starter (Fig. 5b and c), and disappeared throughout aging (Fig. 5e and f). After being bottled for 18 months the microfibrous material of the cell wall appeared more diffuse and the midzone seemed to be degraded, making both fibrous layers fuse (Fig. 5d, black arrow). From the 30th month onward, the walls consisted of a tangled structure of electron-dense fibers (Fig. 5e and f). It was also possible to distinguish dark spots in the walls of cells at 48 months of aging (Fig. 5f). Based on the wall structure schemes proposed by Lipke & Ovalle (1998) and Osumi (1998) and on the fact that uranyl ions react strongly with phosphate and amino groups (Bozzola & Russell, 1999), we believe that these electron-dense spots could be protein deposits.

Figure 5.

High magnification of ultrathin sections of Saccharomyces cerevisiae lees walls harvested at 1 month (a), 9 months (b), 13 months (c), 18 months (d), 30 months (e), and 48 months (f) of aging. Bars: 200 nm.

Given our results we think that cell wall degradation during autolysis in real conditions could be summarized as follows: (1) the amorphous midzone is progressively degraded (Fig. 5), probably due to the activity of glucanases that hydrolyzed glucans, thus releasing mannoproteins such as flocculation proteins (Bony et al., 1997; Klis et al., 2002); (2) residual activities of glucanases released glucans (Alexandre & Guilloux-Benatier, 2006), provoking the inner layer to lose its microfibrous continuous structure (Fig. 5d, e and f); (3) at long-term aging (Great Reserve Cava), the inner and the outer layer of the cell wall formed a tangled structure of fibers which could be composed of β1,3-glucans associated with proteins (Figs 5e, f and 6e, f, g). In spite of these changes, S. cerevisiae cells maintained their integrity throughout the 48 months of wine aging. Our results were comparable to the depletion of the cell wall observed in simulated (Hernawan & Fleet, 1995) and real sparkling wine conditions (Piton et al., 1988), and related to the decrease in the hydrophobicity of wine lees and in cell surface proteins (Vichi et al., 2010), which were associated with the release of yeast cell wall mannoproteins.

Figure 6.

Ultrathin sections of Saccharomyces cerevisiae lees showing different stages of autolysis after 9 months (a–d) and 48 months (e–f) of aging. Cells from the same sample showed different degrees of autolysis (see text for explanation). Bars: (a) 200 nm, (b) 1 μm, (c) 500 nm, (d) 200 nm, (e) 200 nm, (f) 1 μm, (g) 1 μm, (h) 100 nm.

Our results improve on the work of Piton et al. (1988) on the cell wall of lees after long aging periods in enological conditions. These authors worked on champagne bottle-fermented samples and observed that the disappearance of the internal layer is complete after 6 month of bottling. However, in the current study, we observed the changes in the cell wall's internal layer after 18 months (Reserve sparkling wine) (Fig. 5d, e and f). This discrepancy in time could be owing to the yeast strain and/or to the TEM procedure. We have studied the autolysis of lees samples of bottles from three different wine blends produced with the same yeast strain for 4 years, which allowed us to determine the sequence of cell wall degradation throughout the aging process.

Evolution of cell cytoplasm lees

Samples collected during wine aging consisted of lees at different stages of cytoplasmic degradation (Figs 4 and 6). Differences between cells in their cytoplasmic content were also reported by Hernawan & Fleet (1995) in cells undergoing autolysis in model solutions and by Martínez-Rodríguez et al. (2001) in one 12-month-old bottle of sparkling wine. Individual cells in an asynchronous batch culture, such as in sparkling wine bottles, differ in many variables related to the complexity of the yeast cell cycle. Differences within individual cells in their phase in the cell division cycle, their genealogical age, their size, and in clonal variability (Porro et al., 2009; Portell et al., 2011) may yield the cell morphology distribution of the aging lees. In addition, stationary-phase cultures of S. cerevisiae include both quiescent and nonquiescent cells (Allen et al., 2006), increasing the complexity of the yeast population of lees. Thus, we were not able to establish a time sequence of cytoplasmic autolysis during aging as we did for cell wall degradation.

In general, owing to plasmolysis, the thickness of the periplasmic space (PS) increased with time. Furthermore, although the CM formed deep invaginations and some breakages of the membrane were also observed, the double-layered ultrastructure was preserved throughout aging period (Fig. 6a and e). Although S. cerevisiae responds to ethanol stress by increasing fatty acyl chain length and the proportion of unsaturated fatty acids and sterols in the plasma membrane (Bauer & Pretorius, 2000; Walker & Van Dijck, 2006), one of the major consequences of yeast cell exposure to high levels of ethanol is the disruption of membrane structural integrity (Walker & Van Dijck, 2006). This is in agreement with our observation of plasma membrane (CM) invaginations and breakages after 9 months of aging (Fig. 6a and e dashed arrows). It has been reported that plasma membrane degradation starts after 3 to 6 months, with lipids being released into the medium (Alexandre & Guilloux-Benatier, 2006), but little is known about the fate of the plasma membrane during autolysis in sparkling wine aging on lees. Experiments in a model wine system showed that no phospholipids were released into the medium (Hernawan & Fleet, 1995; Pueyo et al., 2000), which could suggest that the plasma membrane is degraded inside the cell. Fragments of double-layered membranes from the plasma membrane or from organelles such as mitochondria, the nucleus, or autophagosomes were observed in the cytoplasm of the cell (Fig. 6c, dashed arrows). Our micrographs may indicate that the first step of plasma membrane degradation during aging on lees is the production of small fragments that remain inside the cell (Fig. 6a). These fragments of bilayered membranes would then be digested.

Autolysis was accompanied by extensive loss and disorganization of the intracellular contents (Figs 4 and 6) (Hernawan & Fleet, 1995; Martínez-Rodríguez et al., 2001). Nonelectron-dense lipid droplets could be observed in lees from 9 months (Fig. 6b) to 48 months (Fig. 6f) of wine aging. Degradation of the cytoplasm content was already observed at 9 months, as vesicles with different material could be detected (Fig. 6c, solid arrows) and ribosomes (R) were present in the periplasmic space (Fig. 6d). However, the nucleus and mitochondria were observed in some cells 48 months old (Fig. 6g). Figure 6h (arrow) shows some double-membrane organelles, but owing to the degradation process it was not possible to discern if they belonged to the Golgi system or the endoplasmic reticulum (Giddings, 2003; Osumi et al., 2006; Yamaguchi et al., 2011). Our results showed that cytoplasmic degradation was a process that extended for at least 4 years. Cell depletion, nevertheless, was not complete after 48 months in the bottle (Figs 4d and 6f, g). These is in agreement with the results reported by Leroy et al. (1990) and Moreno-Arribas et al. (1998) that there is still autolytic activity in yeasts after 60 and 31 months of aging in wine, respectively. These authors based their conclusions on the analysis of the products released into the wine. In addition, a great increase of intracellular proteolytic activity was reported in sparkling wine (Feuillat & Charpentier, 1982) and in champagne (Leroy et al., 1990) during the aging process. Moreover, an increase in the concentration of nucleic acid materials (Leroy et al., 1990) and in the concentration of the most representative nucleotides (Zhao & Fleet, 2003, 2005; Charpentier et al., 2005) was observed in champagnes, which could indicate the degradation of organelles such as the nucleus, mitochondria, and ribosomes, and subsequent degradation to nucleotides (Fig. 6). Our TEM images may agree with analytical data obtained by several authors, among others Leroy et al. (1990), Charpentier et al. (2005), and Moreno-Arribas et al. (1996), which described compositional changes in sparkling wine during aging.

This is the first time that high-resolution TEM images have been obtained, which provides evidence of several morphological changes in the ultrastructure of S. cerevisiae lees throughout real long-term aging. Most of the studies on yeast autolysis have focused on the chemical analysis of wine, but few of them have studied the cytological changes of yeasts during aging. Many of the changes observed in our work could explain the evolution of the compositional and sensorial characteristics of sparkling wines reported by previous authors. Finally, our results suggest that autophagy is involved in the autolytic process. Moreover, different types of autophagy were observed in lees in real enological conditions. Further research should be conducted to understand the role of microautophagy and macroautophagy in the autolysis of lees in sparkling wine during aging.


This study was supported by the Spanish Ministerio de Ciencia y Tecnología, project AGL2008-03392, by the Generalitat de Catalunya, project SGR2009-606, and by XaRTA (Xarxa de Referència en Tecnologia dels Aliments; ajuts de valorització de la recerca), and through a MEC grant to the PhD student J.J. G-C. We are grateful to Freixenet S.A. wineries for sampling. We appreciate Carmen López-Iglesias laboratory assistance in TEM preparations at the Scientific Technical Center at the University of Barcelona.