Food Research International 55 (2014) 83–92

Contents lists available at ScienceDirect

Food Research International

j ourna l homepage: www.e lsev ie r .com/ locate / foodres

Effects of a β-glucanase enzymatic preparation on yeast lysis duringaging of traditional sparkling wines

Sara Torresi a,1, Maria T. Frangipane a,1, Anna M.V. Garzillo b,2, Riccardo Massantini a,3, Marina Contini a,⁎a Department for Innovation in Biological, Agro Food and Forest systems, Via S. Camillo de Lellis, snc, 01100 Viterbo, Italyb Department of Ecological and Biological Sciences, Largo dell'Università, snc, 01100 Viterbo, Italy

⁎ Corresponding author. Tel.: +39 0761 357495; fax: +E-mail addresses: sara_torresi@virgilio.it (S. Torresi), m

(M.T. Frangipane), amg@unitus.it (A.M.V. Garzillo), massamcontini@unitus.it (M. Contini).

1 Tel.: +39 0761 357495; fax: +39 0761 357498.2 Tel.: +39 0761 357316; fax: +39 0761 357498.3 Tel.: +39 0761 357496; fax: +39 0761 357498.

0963-9969/$ – see front matter © 2013 Elsevier Ltd. All rihttp://dx.doi.org/10.1016/j.foodres.2013.10.034

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 July 2013Accepted 21 October 2013

Keywords:On lees agingAutolysisEnzymatic preparationβ-GlucanaseMéthode champenoiseSparkling wine

Scientific researches on characterization of the commercial enological preparation of LallzymeMMX®containingβ-glucanase and its influence on autolysis of different yeast strains typically employed in the production of spar-kling wine are lacking. The aim of the present work was to plug a gap in this field, studying the β-glucanase ac-tivity of LallzymeMMX® and its interactions with BCS103® and EC1118® yeast strains. The results showed thatβ-glucanase was slightly inhibited by ethanol, but its residual activity at wine pHwas sufficient for the purposes.Kinetic parameters showed a better enzyme–substrate complex formation for the EC1118® strain. The influenceon yeast lysis during 12 months of bottle-agingwasmonitored, demonstrating that enzymeaddition did not sub-stantially influence either the content and progression of total proteins, or foam characteristics. However, scan-ning and transmission electronmicroscopy images and free amino acid analysis indicated β-glucanase improvedcell wall degradation of both selected yeasts, evidencing a lower autolytic capacity of the BCS103® strain. Ourstudy demonstrated that addition of β-glucanase catalyzed cell disorganization and promoted release of yeastcomponents into sparkling wine, with strain-dependent effects. Therefore, employment of β-glucanase richLallzyme MMX®might effectively accelerate some aging characteristics of traditional sparkling wines.

© 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The production of sparkling wine by the traditional (champenoise)method is a long process that follows many steps, most of them requir-ing long time and expensive skilled labor (Torresi, Frangipane, & Anelli,2011). It involves the production of the base wine, which undergoes asecond fermentation in the bottle duringwhich CO2 and alcohol contentincrease, followed by aging on yeast lees for several months. The deathand autolysis of yeasts which occur during aging is essential for the finalcharacteristics of the product and improves its organoleptic properties(Pérez-Serradilla & Lauque De Castro, 2008).

Over the years many attempts to simplify this complex production pro-cess have been made, while at the same time trying to keep the qualityand the characteristics unaltered. Some new approaches in this area in-clude the improvement of the yeasts used in bottle secondary fermenta-tion and the design of procedures to accelerate the natural process ofyeast autolysis (Pozo Bayόn, Martínez-Rodríguez, Pueyo, & Moreno-

39 0761 357498.tfrangi@unitus.itnti@unitus.it (R. Massantini),

ghts reserved.

Arribas, 2009). Among the various factors one must consider in the se-lection of strains for second fermentation, their autolytic capacity isone of the most important. In fact, a good autolytic strain will producea better quality sparkling wine compared to a strain in which this char-acteristic is not present (Martínez-Rodríguez, Carrascosa, Barcenilla,Pozo Bayón, & Polo, 2001). Many researchers have addressed their stud-ies towards the selection of autolytic yeast strains for secondary in bot-tle fermentations (Gonzalez, Martínez-Rodríguez, & Carrascosa, 2003;Todd, Fleet, & Henscheke, 2000).

The cell wall of Saccharomyces is made of mannoproteins (high mo-lecular weight polysaccharides conjugated with proteins) crossed byfibers of glucan and chitin (Pretorius, 2000). Its rupture after yeastdeath involves the action of β-glucanases, classified as endo- and exo-glucanases, which can hydrolyze the β-O-glycosidic linkage of β-glucan chains, leading to the release of glucose and oligosaccharides(Dubourdieu, Villetaz, Desplanques, & Ribéreau Gayon, 1981). As a con-sequence of the cell wall structure breakdown, several cytoplasmic andparietal compounds are released into the sparkling wine, which canmodify its organoleptic and foaming properties with positive effectson the product characteristics (Alexandre & Guilloux Benatier, 2006;Pozo Bayόn et al., 2009).

Autolysis needs several months to occur; low wine aging tempera-ture causes a low death rate and low enzymatic reaction rates, thus ac-counting for the slowness of the process. For years on end, twomethodshave been available to accelerate the autolysis during sparkling wineproduction: adding yeast autolysates to the wine and increasing the

84 S. Torresi et al. / Food Research International 55 (2014) 83–92

aging temperature (Charpentier & Feuillat, 1993). However, both tech-niques caused organoleptic defects in the final product, often describedas “toasty”. Lately, someproducers employ commercially available inac-tive yeasts just before the secondary fermentation; however, despitetheir practical use, very few scientific papers report specific resultsabout their real effects on sparkling wine.

The endogenous enzymes of grape, yeasts and other microorganismsusually present inmust andwine are often unable to sufficiently catalyzethe different bio-transformations that occur during winemaking. Theseactivitiesmay, however, be enhanced by the use of exogenous enzymaticpreparations. Exogenous enzymes inwinemaking are extensively used toimprove clarification and the release of varietal compounds, to reducethe formation of ethyl carbamate, and to lower alcohol levels (VanRensburg & Pretorius, 2000). Another important application of commer-cial enzymatic preparations concerns the improvement of the yeast lysis,that may contribute to enhancing the release of intracellular and parietalcompounds from yeast cells, thus improving the quality of winesaged on lees. The products useful for promoting yeast lysis are a mixof β-glucanase and other collateral enzymes such as pectinases, whosepositive effect on the characteristics of red and white wines has beenevidenced by several studies (Masino, Montevecchi, Arfelli, & Antonelli,2008; Palomero et al., 2009).

Yeast autolysis improvement may be a promising tool to enhancesparkling wine quality and reduce production times and costs. Somemanufacturers suggest employing β-glucanase enological preparationsboth in wines and in sparkling wine production (see datasheet andtechnical documents for LallzymeMMX®), however their effective util-ity on sparkling wines is only supposed, because the scientific knowl-edge on their effect on these peculiar wines really falls short.

In a recent study, Rodriguez-Nogales, Fernández-Fernández, andVila-Crespo (2012) found that the addition of an enzyme preparationrich inβ-glucanase seemed to increase the aging characteristics of tradi-tional sparkling wines. In another recent work, the same researchersfound that β-glucanase can be an excellent coadjuvant to enhancethe antioxidant properties of sparkling wines (Rodriguez-Nogales,Fernández-Fernández, Gómez, & Vila-Crespo, 2012). These resultsare encouraging but not conclusive and further studies aimed at inves-tigating the action and effects of β-glucanase on sparkling wines areessential.

To better understand the behavior of β-glucanase and to maximizeits effects and benefits on sparkling wine elaboration, the presentwork focuses on the characterization of a commercial enzymatic prepa-ration (Lallzyme MMX®) rich in this enzyme, and on its applicationin sparkling wine produced with the méthode champenoise. To ourknowledge, studies on the activity of this enzymatic preparationand its effect on different yeast strains are not currently available,oriented at sparkling wine production. Therefore, the β-glucanaseactivity was first studied in different buffers and in model wine solu-tions. Both pH and ethanol content effects were also considered, atthe typical values of base wine and sparkling wine. Moreover, interac-tions between the enzymeand twodifferent yeast strains (Saccharomycescerevisiae BCS103® and EC1118®), suitable for the second fermentationof sparkling wines, were evaluated. Subsequently, sparkling wines wereproduced employing these two strains, both in the presence and absenceof the enological preparation, and the influence on yeasts autolysisprocess during 12 months of bottle aging was investigated, by direct(scanning and transmission electron microscopy) and indirect (freeamino acids, total proteins, foam height and stability)methodologicalapproaches.

2. Materials and methods

2.1. Enzymatic preparation stock solution

The enological enzymatic preparation used for all experiments wasLallzyme MMX® (Lallemand Inc., Canada), sourced from Trichoderma

spp. and Aspergillus niger. In the technical data sheet provided by themanufacturer, β-glucanase activity is not reported. Other declared collat-eral enzymatic activities are: polygalacturonase, 1840 IU g−1; pectin-esterase, 545 IU g−1; pectin-lyase, 25 IU g−1. A stock solution ofLallzyme MMX® at 1 mg ml−1 was prepared daily in deionized water.

2.2. Yeast stock solution

The two different yeast strains used were: S. cerevisiae BCS103®(Fermentis, Division of S.I. Lesaffre, France) and S. cerevisiae LalvinEC1118® (Lallemand Inc., Canada). A fresh stock solution of eachyeast at 1 mg ml−1 was prepared daily in deionized water.

2.3. β-Glucanase substrate stock solution

The selected substrate for β-glucanase enzymatic activity determi-nations was laminarin (Humbert Goffard et al., 2004; VázquezGarcidueñas, LealMorales, & Herrera Estrella, 1998), a β(1 → 3)-glucanfrom Laminaria digitata. A stock solution of laminarin at 1 mg ml−1 wasfreshly prepared daily in deionized water.

2.4. β-Glucanase enzymatic activity evaluation

The enzymatic assays were performed incubating at 25 °C the sub-strate laminarin stock solution (0.4 ml), LallzymeMMX® stock solution(0.08 ml) and 5.92 ml ofMcIlvaine buffer (McIlvaine, 1921), both in thepresence (1.6 ml of the stock solution) and the absence (1.6 ml of de-ionized water) of yeast. Blank assays without enzyme or Laminarinwere also performed. After 45 min of incubation, the reaction wasstopped by heat inactivation (100 °C, 2 min) (Humbert Goffard et al.,2004). The glucose released was determined after centrifugation,through the enzymatic test combination D-glucose/D-fructose kit(R Biopharm, Germany), following the instructions of the manufactur-er. The test is based on the NADPH amount formation that is stoichiomet-ric with the amount of D-glucose; NADPH was measured by absorbanceincrease at 340 nm by using a Perkin Elmer Lambda 25 UV/VIS spectro-photometer (Beaconsfield, UK). Since the laminarin substrate is a poly-saccharide having a variable molecular weight, the enzymatic activitywas expressed as the amount of enzyme that releases 1 mg l−1 min−1

of glucose instead of International Unit (μmol min−1 of laminarin); sub-strate concentration was expressed as mg ml−1 instead of mM. Enzymeactivity (U ml−1) was then determined as described in themanufacturer's instructions, adequately modified as follows:

U ml−1 ¼ ΔA=incubationtime minð Þ � 180:16 g mol−1� �

� finalvolume mlð Þh i

=6:3

mM−1 cm−1� �

� samplevolume mlð Þ

where:ΔAwas obtained by subtracting fromeach sample the absorbanceof the blank assays; 180.16 (g mol−1) is the D-glucose molecular weight;6.3 (mM−1 cm−1) is the extinction coefficient of NADPH at 340 nm.

Moreover, the enzyme activity per milligram of total proteins(Specific Activity) was evaluated at maximal pH, and expressed asmgglucose l−1 min−1 mgprotein−1 .

2.5. Effect of pH on β-glucanase enzymatic activity

The effect of pH on the β-glucanase activity of LallzymeMMX®wasdetermined in McIlvaine buffer (McIlvaine, 1921), both in the presenceand the absence of yeast cells (EC1118®or BCS103® strains). This buffersolution was employed for all the experiments because of its capacity tocover a pH range from 2.2 to 8.0, thus avoiding drawbacks caused by theuse of different buffer solutions. Samples were tested at different pH(2.60, 3.20, 4.00, 5.00, 6.00 and7.00), and the enzymeactivitywas deter-mined as described above.

85S. Torresi et al. / Food Research International 55 (2014) 83–92

2.6. Effect of ethanol content on β-glucanase enzymatic activity

The ethanol influence on β-glucanase enzymatic activity of LallzymeMMX® was analyzed both in the presence and the absence of yeast.Two different ethanol concentrations were selected for the experi-ments, 11.3% (v/v) and 12.7% (v/v), corresponding to those of a basewine and a sparkling wine respectively. The samples were prepared asdescribed above, but replacing the McIlvaine buffer with the tartaricone (0.1 MatwinepH, 3.20). As control, an experimentwithout ethanolwas also performed. Enzyme activity was determined as describedabove.

2.7. Enzyme kinetics and kinetic parameters determination

Enzyme kinetics were performed in theMcIlvaine buffer at the opti-mal pH value, obtained from the activity vs. pH curves. LallzymeMMX®stock solution (0.08 ml) and increasing amounts (from 0 to 1 ml) ofsubstrate laminarin stock solution were employed. Analogous experi-ments were prepared in the presence of EC1118® or BCS103® yeaststrains. The β-glucanase activity was evaluated as described above.The following kinetic parameterswere determined by the non-linear re-gression analysis of the Michaelis–Menten kinetic equation, employingthe software GraphPad Prism 4®:

Vmax corresponding to the maximum velocity at which an enzymecatalyzes a reaction, expressed as mgglucose l−1 min−1;

KM representing the amount of substrate corresponding to 1/2Vmax. This parameter reflects the efficiency of the enzyme–substrate complex formation and was expressed as mg ml−1

of substrate.

2.8. Sparkling wine production

Sparkling wines were manufactured industrially by the Falescowinery in Montecchio (Terni, Italy), following the traditionalmethod (secondary fermentation in the bottle). The base wine wasproduced with Roscetto grapes, an autochthonous white grape varietybelonging to the Verdicchio family and widespread in the Umbria andLatium territories (Italy), following the classic white wine technology.The principal enological parameters of the base wine were the fol-lowing: pH, 3.2; alcohol, 11.3% v/v; total acidity, 6.8 g/l of tartaricacid; total SO2, 40 mg/l; free SO2, 12 mg/l. Like Martínez-Rodríguez,Carrascosa, Martín Àlvarez, Moreno-Arribas, and Polo (2002), we didnot add bentonite to the base wine, so as to avoid interferences withenzyme activity and the nitrogen compounds released during yeastautolysis.

A pied de cuve containing EC1118® or BCS103® yeast strains(0.2 g l−1) was used to inoculate the base wine. 80 bottles were pre-pared with yeasts alone (40 bottles for each strain), and 80 withLallzyme MMX® at final concentration of 0.01 g l−1 (40 bottles foreach strain). The bottles were then stacked into caged bins andstored in a special place furnished for the setting of the sparkle. Thebottle storage area was cool and had minimum temperature fluctua-tions andminimum lighting. Secondary fermentation and aging withyeasts were carried out at 14 °C. Three bottles for each experiment(both in the absence and the presence of Lallzyme MMX®) werefitted with an aphrometer, to measure the CO2 pressure increasewithin the bottle. Samples for analyses were taken at 0 (base wine)and at 4, 8 and 12 months after tirage. At the end of the observationtime, all sparkling wines reached a pH value of 3.5 and an alcoholcontent of 12.7% v/v.

All the analyses were performed on vacuum degassed wines ob-tained by mixing the content of 2 bottles.

2.9. Direct evaluation of yeast lysis in sparkling wine

A direct estimation of yeast lysis was obtained on sparkling winesamples by SEM (scanning electron microscopy) and TEM (transmis-sion electron microscopy) observations of yeast lees, performed by theCIME (Interdepartmental Centre of Electron Microscopy, University ofTuscia, Viterbo). Vacuum degassed samples were centrifuged, the pelletcollected and immediately prepared for SEM and TEMobservations. Themethod of Wright (2000) was used for sample preparation.

2.10. Indirect evaluation of yeast lysis in sparkling wine

2.10.1. Total proteinsTotal protein content was determined by the Bradford dye binding

assay (Bradford, 1976). Briefly, 100 μl of wine sample were introducedinto a 4 ml cuvette, then 3 ml of Bradford reagent were added. After45 min of incubation, the absorbancewasmeasured at 595 nm. The cal-ibration curve was prepared in a Tartaric buffer at pH 3.20, using BSA(bovine serum albumin) 2 mg ml−1 as standard stock solution.

2.10.2. Free amino acid analysisFree amino acid analyses were carried out by the Laboratory of Pro-

teomics andMass Spectrometry of the University of Tuscia. For the sam-ple preparation, solid phase extraction using Extra-Sep C18 columns(Lida Manufacturing Corp., 25th Avenue, Kenosha, WI 53143-6615,USA)was performed; the obtained samplewas dehydrated and then re-suspended into solvent A (0.1% formic acid in water). An Ultimate 3000rapid resolution fast HPLC system (LC Packings, Dionex, Sunnyvale,USA) equipped with a Dionex Acclaim RSLC 120 C18 column(2.1 mm × 150 mm, 2.2 μm) was used to separate the extracted me-tabolites. LC parameters: injection volume, 20 μl; column temperature,30 °C; flow rate, 0.2 ml min−1. The LC solvent gradient and timetablewere identical during the whole period of the analysis. A 0–95% lineargradient of solvent A (0.1% formic acid in water) to B (0.1% formic acidin acetonitrile) was employed over 15 min, followed by solvent B holdof 2 min, returning to 100% A in 2 min, and a 6 min post time solventA hold. The metabolites were directly eluted into a High Capacity ionTrap HCTplus (Bruker Daltonik, Bremen, Germany). Mass spectra formetabolite extracted samples were acquired in negative ion mode. ESIcapillary voltage was set at 3000 V (−) ion mode. The liquid nebulizerwas set at 30 psig and the nitrogen drying gas at a flow rate of9 L min−1.

2.10.3. Foaming propertiesFoaming properties were evaluated following the method of Pueyo,

Martín Álvarez, and Polo (1995), suitably modified. The equipmentconsisted of a graduated glass cylinder (530 × 15 mm) (SteroglassS.R.L., San Martino in Campo, Perugia), with a glass frit at the bottomof the column (pore size 10 μm). The column was attached to a massflow meter and controller EL-Flow Select model F201CV-200-AGD-33-Z-MFC (Bronkhorst Hi-Tech; Ruurlo, Netherlands) connected to a CO2

fitting. Instrument control and CO2 flux setting was obtained by theSoftware Flow-Bus vs. 4.6 provided by Bronkhorst Hi-Tech. Beforeeach measurement, the glass cylinder was accurately washed threetimes with 10 ml of 95% ethanol, three times with 10 ml of Milli-Qwater and finally with 50 ml of sample. Ten ml of degassed sparklingwine were introduced into the glass cylinder, and then a 75 ml min−1

of CO2 flux was introduced into the column through the glass frit. Twoparameters were measured: the capacity to form foam (foam height),i.e., the maximum height (cm) reached by the foam column into theglass cylinder after CO2 injection, and the foam stability, i.e., the time(seconds) elapsed before all the bubbles disappear after CO2 fluxinterruption.

86 S. Torresi et al. / Food Research International 55 (2014) 83–92

2.11. Statistical analysis

All measurements were carried out at least in triplicate (n = 3) andthe results reported as means ± SD. The significance between meanswas evaluated by one-way variance analysis (ANOVA), performedusing the SPSS version 15 statistical package for Windows® (SPSS inc.,Chicago, IL, USA).

3. Results and discussion

3.1. Characterization of Lallzyme MMX® enological preparation

3.1.1. Effect of pH on β-glucanase enzyme activityThe influence of pH on enzyme activity of β-glucanase contained in

LallzymeMMX®was determined, as well as its lytic activity at differentpH on two different yeast strains. The first yeast we employed was S.cerevisiae EC1118®, one of the commercial strains most commonlyused for the refermentation of base wine in sparkling wine production,which presents a good fermentation vigor, excellent alcohol tolerance(up to 18% v/v), capacity to ferment at low temperature, good floccula-tion ability, and good autolytic capacity during sparkling wine aging,leading to the release of autolytic products into the extracellular envi-ronment having positive effects on foaming properties and sensorycharacteristics of sparkling wines (Martínez-Rodríguez, Carrascosa, &Polo, 2001; Martínez-Rodríguez et al., 2001). The second yeast wasS. cerevisiae BCS103®, a strain selected in the Champagne region onChardonnay for its excellent fermentation characteristics and its greatresistance to extreme wine making conditions (information providedby the manufacturer, Technical Data Sheet Lalvin).

According to other authors (De Marco & Felix, 2007; Noronha &Ulhoa, 2000), who studied β-glucanases produced by Trichodermaharzianum, the maximal activity of β-glucanase contained in LallzymeMMX® was detected at pH 5.0 (Table 1). However, a higher activityfor both samples containing yeasts was evidenced by the statisticalanalysis (p ≤ 0.01) at 4.0 and 5.0 pH values, with no statistically signif-icant differences between the two tested strains. In particular, 130% and134% of relative activity was found in samples containing BCS103® andEC1118® respectively, at maximal pH. At wine pH (3.2), the relative ac-tivity with respect to themaximumwas 66% (without yeast), 70% (withBCS103® strain) and 71% (with EC1118® strain), indicating that the re-sidual activity was useful for the purposes of this study, even if wine pHis not the optimum value for β-glucanase. Remarkably, the enhance-ment of enzyme activity due to the yeast presencewas about 10% higherat wine pH than at the maximum; moreover, at pH 3.2, EC1118® sam-ple showed a better activity (p ≤ 0.01) than BCS103® one.

The results showed the effectiveness of the enzymatic preparation inlysis enhancement of both selected yeasts, especially the EC1118®strain, leading to a consistent release of the intracellular yeast enzymesand more efficient cell lysis, more evident at wine pH. No activity was

Table 1Activity (mean ± SD) and relative activity at different pH of β-glucanase contained in Lallzym

Enzyme activity (mgglucose l−1 min−1)

pH Ea BCb ECc

2.6 0.043 ± 0.001 a 0.046 ± 0.002 a 0.053 ± 0.003 a3.2 0.106 ± 0.002 a 0.147 ± 0.001 b 0.153 ± 0.002 c4.0 0.153 ± 0.007 a 0.193 ± 0.004 b 0.183 ± 0.003 b5.0 0.161 ± 0.002 a 0.209 ± 0.004 b 0.215 ± 0.003 b6.0 0.117 ± 0.004 a 0.115 ± 0.003 a 0.116 ± 0.004 a7.0 0.043 ± 0.001 a 0.048 ± 0.002 ab 0.051 ± 0.002 b

Within the same line, values with different letters indicate statistically significant differences (a Enzymatic preparation alone.b Enzymatic preparation plus BCS103® yeast strain.c Enzymatic preparation plus EC1118® yeast strain.d Relative activity of E, EC and BC samples, compared to that manifested at pH 5.0.e Relative activity of EC and BC samples, compared to E samples.

detected in the control samples containing yeasts and substratewithoutLallzymeMMX®, thus demonstrating that the activity noticed is due tothe Lallzyme MMX® β-glucanase hydrolysis of yeast cell wall.

In the enzymatic preparation used for all experiments, 1.5% of totalprotein contentwas detected. This datum allows us to calculate the Spe-cific Activity (SA) value, a parameter which gives an estimate of the pu-rity of the enzyme. The SAwe found for LallzymeMMX®β-glucanase atmaximal pH was 0.107 ± 0.003 mgglucose l−1 min−1 mgprotein−1 .

3.1.2. Effect of ethanol content on enzyme activityCommercial glucanase preparations are believed not to be influ-

enced by alcohol content up to 14% v/v (Van Rensburg & Pretorius,2000). However, ethanol influencewas evidenced inβ-glucanase activityof the enzyme preparation VinoflowG (Laffort Oenologie), a commercialblend active in promoting hydrolysis of β-glucans and degradation ofpectic substances (Zinnai, Venturi, Quartacci, Andreotti, & Andrich,2010). The authors found a decrease in activity of 50% in a model winesolution containing 13% v/v of ethanol at pH 3.3.

To our knowledge, there are no research works reporting the etha-nol effects on LallzymeMMX® β-glucanase activity. Therefore, we test-ed the β-glucanase activity of this commercial preparation at twodifferent ethanol concentrations (11.3 and 12.7% v/v, corresponding tothose of a base wine and a sparklingwine respectively), in the presenceor in the absence of both EC1118® and BCS103® yeast strains. The trialswere carried out in a model wine solution, that is in tartaric bufferpH 3.20 (which is the most similar to real wines), even if no significantdifferences were detected between experiments performed in tartaricor McIlvaine buffers at this pH value (data not shown).

Experimental results, shown in Table 2, demonstrated that additionof ethanol inhibited β-glucanase activity, and that the enzyme activitywith respect to the sample without ethanol was at the same extent forthe two tested alcohol concentrations, both in the absence and the pres-ence of yeasts.Moreover, thefindings showed that the yeasts had a pos-itive effect on the enzyme activity. No significant difference (p ≤ 0.01)was detected between the two tested strains at 11.3 and 12.7 alcoholcontent, whilewithout ethanol EC1118® strain sample showed a higheractivity (p ≤ 0.01) than did BCS103®.

EC1118® yeast showed a linear decrease of activity in the presenceof ethanol, and thus it was possible to extrapolate the E50 (i.e., the 50%of maximum activity), corresponding to an ethanol concentration of20% v/v, a higher value than those generally present in wines. For bothsamples without yeast and with the BCS103® strain instead, the de-crease in enzyme activity was not linearly dependent on tested ethanolconcentrations, but a similar E50 value can be extrapolated.

3.1.3. β-Glucanase enzyme kineticsThe kinetic parameters of β-glucanase contained in Lallzyme

MMX®, both in the presence and the absence of yeast strains, were es-timates at maximal pH (5.0). The kinetic curves we obtained always

e MMX®, in the presence or in the absence of yeast.

Relative activity (%)

Ed BCd ECd BC vs. Ee EC vs. Ee

27 21 25 108 12366 70 71 138 14496 93 85 126 120

100 100 100 130 13473 55 54 98 9927 23 24 111 119

p ≤ 0.01).

Table 2Effect of different ethanol concentrations on Lallzyme MMX® β-glucanase activity(mean ± SD), measured in tartaric buffer (pH 3.20) in the absence or in the presence ofyeast.

Sample Ethanol (% v/v) Enzime activitya

(mgglucose l−1 min−1)Relative activity(% of the maximum)

Eb 0 0.117 ± 0.003 c 10011.3 0.072 ± 0.004 d 6212.7 0.072 ± 0.002 d 62

BCc 0 0.156 ± 0.001 b 10011.3 0.107 ± 0.003 c 6812.7 0.107 ± 0.002 c 68

ECd 0 0.172 ± 0.002 a 10011.3 0.122 ± 0.007 c 7112.7 0.118 ± 0.005 c 69

a Within the column, different letters indicate statistically significant differences(p ≤ 0.01).

b Enzymatic preparation alone.c Enzymatic preparation plus BCS103® yeast strain.d Enzymatic preparation plus EC1118® yeast strain.

87S. Torresi et al. / Food Research International 55 (2014) 83–92

showed a hyperbolic trend, so the elaboration was performed using theMichaelis–Menten equation.

No β-glucanase activity was detected in samples containing yeastsand substrate, without enzyme. The Vmax (mgglucose l−1 min−1) forthe Lallzyme MMX® β-glucanase (without yeasts) was 0.181 ± 0.003.The samples plus yeasts showed Vmax values (0.244 ± 0.007 and0.259 ± 0.010, in the presence of EC1118® or BCS103® strain, respec-tively) statistically higher (p ≤ 0.01) thanwith the enzyme alone. How-ever, significant differences (p ≤ 0.01) between the two tested strainsdid not emerge.

The efficacy of EC1118® yeast was confirmed by the KM value(0.148 ± 0.013 mg ml−1 of substrate) found in its presence, which re-sulted significantly lower (p ≤ 0.01) than that detected in the presenceof BCS103® strain (0.257 ± 0.0024) and closer to that of the enzymewithout yeasts (0.134 ± 0.008), indicating a better enzyme–substratecomplex formation.

3.2. Evolution of the secondary fermentation in sparkling wine samples

All sparklingwines produced completed the secondary fermentationin 42 days, as indicated by internal bottle pressure monitoring. The sec-ondary fermentation started 2 days later for sparkling wines producedwith EC1118® strain, both in the absence and the presence of LallzymeMMX® (data not shown). However, internal bottle pressure becamesimilar for all the wines at 8 days and then constantly increased, morerapidly for samples produced with EC1118® strain both in the absenceand the presence of Lallzyme MMX®. At the end of fermentation(42 days), all the sparkling wines reached the same pressure value of6 bar.

3.3. Monitoring of yeast lysis during over lees aging of sparklingwine samples

The effect of Lallzyme MMX® on yeast lytic process was monitoredby both SEM and TEM observations of yeast cells, which reveal ultra-structural morphological changes occurring during aging (direct evalu-ation), and determining the modifications in total protein and freeamino acid profile occurring in sparkling wine samples as a conse-quence of yeast lysis (indirect evaluation).

3.3.1. Direct evaluation of yeast lysis during aging of sparkling wineMany studies aimed at analyzing the products released during autol-

ysis are available, while less attention has been paid for evidencing theultrastructural changes occurring in yeast cells (Martínez-Rodríguez,Polo, & Carrascosa, 2001). To our knowledge, the present work is thefirst in which ultrastructural morphological changes of yeasts are

described, during sparkling wine on lees aging process, both carriedout in the presence of cellulolytic enzymes.

In Fig. 1, EC1118® and BCS103® yeast cell images evidenced that4 months after tirage an internal cytoplasmic disorganization waspresent, as shown by the TEM images obtained both in the absence(control) and the presence of Lallzyme MMX® (Fig. 1b, left) respectto time 0 (Fig. 1a). SEM images (Fig. 1b, right) instead showed no differ-ences 4 months after tirage respect to time 0. As mentioned above, celllysis is a slow process, especially under the restrictive conditions inwhich secondary fermentations occur.

Eight months after tirage, cytoplasmic disorganization was moreevident and the cell wall started collapsing in the presence of LallzymeMMX® (Fig. 1c). SEM images (Fig. 1c, right) showed a slight roughcell surface, indicating autolysis had begun; this process was moremarked for EC1118® yeasts and cells lysis was slightly more evidentin sparkling wines produced with Lallzyme MMX® addition. Similarfindings were obtained by TEM images (Fig. 1c, left), showing thatLallzyme MMX® addition enhanced the cell wall degradation.

At the end of sparkling wines aging process (12 months), SEM ob-servations showed that yeast cells were collapsed, showing smallersize than time 0, both in the presence and the absence of LallzymeMMX® (Fig. 1d, right). SEM images revealed a rough cells surface; thecell wall was unbroken, but presented many ridges or folds, both inthe presence and the absence of β-glucanase, indicating cell lysis. TEMobservations at 12 months (Fig. 1d, left) showed a clear cytoplasmicwrinkling and a membrane detachment from cell wall for both yeasts.Similar results were observed by Martínez-Rodríguez et al. (2001)after 12 months of traditional sparklingwines aging, using low temper-ature SEM. Moreover, the authors found that in the fractured cells agreat part of their cytoplasmic contentwas not solubilized but remainedinside the cell, indicating that natural autolysis of yeast is a long-lastingprocess, which continues for at least 12 months.

3.3.2. Indirect evaluation of yeast lysis during aging of sparkling wine

3.3.2.1. Total protein content. Even if proteins are minor constituents ofwine, they contribute to its quality. In fact, proteins inwines are respon-sible for the “body” sensation, can bind volatile compounds retainingwine aroma, and also have a positive effect on foam stability(Brissonnet & Maujean, 1991; Luguera, Moreno-Arribas, Pueyo,Bartolomé, & Polo, 1998). Proteins seem to be foam active substancesbecause of their surface properties. They act as tensioactive compoundsand enhance foam stability (García, Aleixandre, Álvarez, & Lizama,2009). Moreover, Rodriguez-Nogales et al. (2012) found that total pro-teins were highly and positively correlated with antioxidant propertiesof traditional sparkling wines. During the autolysis process, proteaseshydrolyze yeast proteins to lower molecular weight compounds.

Total protein content was monitored using the Bradford Method.Despite the method has some limitations (Marchal, Seguin, &Maujean, 1997), it is used frequently in enology because of its rapidityand its reproducibility. This method was adopted by other authors forestimating the evolution of total protein during yeast lysis in traditionalsparkling wines (Martínez-Rodríguez et al., 2002; Rodriguez-Nogaleset al., 2012).

Total protein showed a decline 4 months later (Fig. 2). A decrease insparkling wine total proteins at 3 month aging because of their insolu-bilization due to alcohol content increase and their partial hydrolysiswas observed by other authors (Nuñez, Carrascosa, González, Polo, &Martínez-Rodríguez, 2005; Rodriguez-Nogales et al., 2012). Similarlyto the results of Rodriguez-Nogales et al. (2012) and Martínez-Rodríguez et al. (2002), an increment was instead detected during thenext period of aging, as a consequence of yeast autolysis that releasesprotein and polypeptide into the wine (Alexandre & Guilloux Benatier,2006; Pérez-Serradilla & Lauque De Castro, 2008).

Total protein increase was particularly evident from 8 to 12 months,especially for the samples produced with BCS103® yeast. In sparkling

Fig. 1. TEM (on the left), and SEM (on the right) images of BCS103® and EC1118® yeast cells at: time 0 (a) and after 4 months (b), 8 months (c) and 12 months (d) from tirage, in theabsence (EC1118; BCS103) or in the presence (EC1118 + β-gluc.; BCS103 + β-gluc.) of Lallzyme MMX®.

88 S. Torresi et al. / Food Research International 55 (2014) 83–92

wines obtained with the same strain, total protein content of samplestreated with the enzyme was slightly higher, but the differences werenot found to be statistically significant (p ≤ 0.01). It follows that theemployment of enzymedid not substantiallymodify either the progres-sion or the content of sparkling wine total protein, which at the end ofaging time was found to depend only on the yeast strain.

Statistically significant differences (p ≤ 0.01) did not emerge in totalprotein values of samples aged 4 and 8 months, even if 8 months aftertirage SEM and TEM images clearly indicated that autolysis had begun,and that the process was more evident for EC1118® yeasts and slightly

more evident in sparkling wines produced with Lallzyme MMX® addi-tion (Fig. 1). Hence, we can deduce that total protein evolution evaluatedwith the BradfordMethodwas not an efficientmarker formonitoring thebeginning of the autolysis process.

3.3.2.2. Free amino acid profile.Amino acids are considered goodmarkersfor following autolysis and have an impact on the quality of tradi-tional sparkling wines, being precursors of some aroma compounds(Alexandre & Guilloux Benatier, 2006). Therefore, the free amino acidic

Fig. 2. Total proteins of the basewine (time 0) and sparklingwines producedwith EC1118®and BCS103® yeast strains, in the presence or in the absence or of Lallzyme MMX®, at 4, 8and 12 months after tirage.

89S. Torresi et al. / Food Research International 55 (2014) 83–92

profile and their total content in base wine and sparkling wines duringaging were monitored, and the results shown in Table 3.

The amino acid profiles of the basewine and the sparklingwines weobtained were quite different from those reported by Martínez-Rodríguez et al. (2002), who observed noticeable variation in the freeamino acid content of sparklingwine during autolysis, which dependednot only on the aging time but also on the yeast strain used for re-fermentation. These authors evidenced four main stages in the aging ofsparkling wines on lees: during the second fermentation (first stage),amino acids and proteins diminished and peptides were released; in thesecond stage, in which viable cells coexist with dead cells, a release ofamino acids and proteins occurred, and peptides were degraded. In thethird stage, no viable cells were present and the release of both proteinsand peptides predominated. In the last phase, which occurred 9 monthsafter tirage, the authors found a decrease of amino acid concentration. Asimilar trendwas observed in the free amino acid content of the sparklingwine samples we studied.

In our samples, the most abundant free amino acid found at time 0(base wine), was lysine (7,84 mmol l−1), which made up 74.0% oftotal free amino acids, followed by GABA (γ-aminobutyric acid, 7.2%);some amino acids, such as glutamine, valine and isoleucine, were notdetected at all.

Four months later, a strong decrease of the free amino acid sumwasdetected in all sparkling wines, which became about 20–30% of the ini-tial value for all samples, probably due to their assimilation by the yeastsfor protein synthesis and due to their accumulation into vacuoles as areserve. Lysine, which was the markedly prevalent free amino acid inbase wine, disappeared; other amino acids, such as serine, alanine, aspar-agine, threonine andmethionine, decreased conspicuously or even disap-peared. However, declinewas not generalized, because some amino acids(i.e., arginine, isoleucine and tryptophan) increased in all samples.

Eight months after tirage, a noticeable augment in free amino acidsum was only detected for EC1118® samples, both in the absence andthe presence of β-glucanase, indicating that autolysis had started(Nuñez et al., 2005). In particular, total free amino acid enhancementwith respect to the previous observation (4 months) was 8.0 timesgreater in EC1118® wine without enzyme and 8.6 times in the corre-spondent sample with the enzyme. Notably, lysine increased fromzero (4 months) to about 90% of total free amino acids. Nitrogen com-pound release is thought to reflect the autolytic activity of yeast, espe-cially the proteolytic activity (Alexandre & Guilloux Benatier, 2006). Atthis stage, BCS103® samples (with and without enzyme) showed aslight increase (not statistically significant, p ≤ 0.01) in the aminoacid sum, indicating that autolysis was not yet evident.

At 12 months of aging, EC1118® samples manifested a noticeabledecrease in the amino acid sum, probably due to deamination, decar-boxylation and/or synthesis reactions which lead to the formation of anumber of traditional sparkling wine aroma compounds (Alexandre &Guilloux Benatier, 2006). The decline resulted particularly intense inthe sample with β-glucanase. In fact, total free amino acids werefound to be about 14% and 60% of the value manifested at 8 monthsfor EC1118® samples with and without enzyme, respectively. In addi-tion, in the presence of the enzyme, lysine disappeared, while in itsabsence, over 60% of this amino acid was retained. More intensediminishing in free amino acids may point to the fact that the yeastswith enzyme had reached a more advanced autolysis stage. In BCS103®yeast, symptoms of autolysis were only evident at this last observationtime, and solely in the presence of β-glucanase, as indicated by theamino acid sum increase up to 25.73 mmol l−1. At this stage, inBCS103®wines with the enzyme, very high amounts of lysine appeared,constituting nearly 88% of total free amino acids; on the contrary, in thesame sample without the enzyme, lysine level remained at zero.

We can note that total free amino acid value and lysine content ofthe BCS103® sample with the enzyme at 12 month aging were analo-gous to that manifested by the EC1118® samples at 8 months (withor without enzyme), as emerged from the statistical analysis whichdid not evidence significant difference among samples (p ≤ 0.01).Thisfinding confirms the previous observations indicating a lower auto-lytic efficiency of the BCS103® strain.

No significant modification (p ≤ 0.01) in total free amino acids wasdetected in BCS103® samples without enzyme, during aging from 4 upto 12 months; this finding indicated that total free amino acid evolutionwas not a goodmarker for monitoring natural lysis process in this yeaststrain.

GABA is not a constituent of proteins, but Saccharomyces can utilize itas a carbon source (Bisson, 1991). This amino acid was found at0.76 mmol l−1 in the base wine (time 0) and around 0.6–0.7 mmol l−1

in all sparkling wine samples until 4 months. At 8 and 12 months,GABA became very reduced or even disappeared in all the samples with-out Lallzyme MMX®, indicating that it was accumulated by yeasts in theintracellular pool or utilized in starvation conditions. On the contrary,after 8 months of aging in the presence of enzyme, GABA levelwas slight-ly increased with respect to the previous observation (4 months), andshowed the same content (0.85 mmol l−1) in samples aged 12 months,independently of the strain used. GABA release into the extracellularenvironment is a consequence of cell lysis. During induced autolysisin a model wine system, GABA increase was observed by Martínez-Rodríguez et al. (2001), too. The low content of GABA in sparklingwine samples without enzyme and its very high levels in samplesadded with enzyme, indicated that β-glucanase improved degradationof the cell wall of both yeast strains studied.

The findings demonstrated the positive effect of β-glucanase on theprocess of cell lysis, especially for EC1118® strain provided with highautolytic power; however, our results suggested that exogenous β-glucanase employmentmay also be helpful in improving amino acid re-lease when strains having minor lytic capacity, such as BCS103®, areutilized.

3.3.2.3. Foaming properties. It is well known that foam properties are im-portant quality characteristics of sparkling wines. For this reason, weanalyzed the evolution of foam properties (height and stability) in thebasewine and in the sparklingwines during aging, with the aim of eval-uating if the addition of β-glucanase could influence this importantqualitative character. The results are shown in Fig. 3.

Up to 4 months after tirage, foam height of EC1118® samplesremained constant, both in the presence and the absence of β-glucanase, while it manifested a decrease in BCS103® samples withor without enzyme (Fig. 3a). However, at 8 month aging, foamheight of BCS103® samples greatly dropped to values not signifi-cantly different (p ≤ 0.01) to those of other wines. At the end of

Table 3Free amino acid content (mmol l−1) in base wine and sparkling wines after 4, 8 and 12 months of aging.

Sparkling wines

BCa ECb

Base wine Alone + Enzymec Alone + Enzymec

Months 0 4 8 12 4 8 12 4 8 12 4 8 12

AALys 7.84 ± 0.42 0 0 0 0 0 22.54 ± 0.90 0 22.40 ± 0.91 13.60 ± 0.60 0 22.5 ± 0.84 0Ser 0.06 ± 0.002 0 0.04 ± 0.001 0 0 0 0.04 ± 0.001 0 0.03 ± 0.001 0.03 ± 0.001 0.02 ± 0.001 0.04 ± 0.001 0.03 ± 0.001Pro 0.07 ± 0.003 0.03 ± 0.001 0.17 ± 0.007 0.09 ± 0.004 0.20 ± 0.011 0.22 ± 0.010 0 0.16 ± 0.009 0.07 ± 0.003 0.03 ± 0.001 0.12 ± 0.006 0 0.10 ± 0.004Cys 0.05 ± 0.001 0.01 ± 0.000 0.03 ± 0.001 0.05 ± 0.002 0.04 ± 0.002 0.04 ± 0.001 0.03 ± 0.001 0.60 ± 0.012 0 0 0.04 ± 0.001 0.03 ± 0.001 0.03 ± 0.001GABAd 0.76 ± 0.024 0.59 ± 0.020 0 0.02 ± 0.001 0.65 ± 0.023 0.70 ± 0.024 0.85 ± 0.024 0.75 ± 0.024 0.15 ± 0.004 0 0.72 ± 0.024 0.85 ± 0.028 0.85 ± 0.030Ala 0.37 ± 0.018 0.03 ± 0.001 0.06 ± 0.003 0 0.05 ± 0.002 0.06 ± 0.002 0.05 ± 0.002 0.06 ± 0.003 0.55 ± 0.030 0.42 ± 0.021 0.06 ± 0.003 0.05 ± 0.002 0.06 ± 0.003Gly 0.07 ± 0.002 0.32 ± 0.019 0.03 ± 0.001 0.60 ± 0.035 0.55 ± 0.030 0.64 ± 0.030 0.04 ± 0.002 0.61 ± 0.022 0.04 ± 0.001 0.03 ± 0.001 0.02 ± 0.001 0.04 ± 0.001 0.04 ± 0.002Asn 0.22 ± 0.008 0.01 ± 0.000 0 0.01 ± 0.001 0.01 ± 0.001 0 0 0.08 ± 0.004 0 0.09 ± 0.004 0 0 0Gln 0 0.03 ± 0.001 0.04 ± 0.001 0 0 0.03 ± 0.001 0.04 ± 0.002 0 0.03 ± 0.001 0.03 ± 0.001 0.03 ± 0.001 0.04 ± 0.001 0.04 ± 0.001Hys 0.03 ± 0.001 0.03 ± 0.001 0.47 ± 0.021 0.05 ± 0.001 0.05 ± 0.002 0.06 ± 0.002 0.42 ± 0.022 0.05 ± 0.002 0.41 ± 0.010 0.22 ± 0.012 0.40 ± 0.020 0.45 ± 0.023 0.47 ± 0.027Thr 0.38 ± 0.018 0.03 ± 0.001 0.09 ± 0.003 0.03 ± 0.001 0.02 ± 0.001 0.02 ± 0.001 0.08 ± 0.002 0.02 ± 0.001 0.08 ± 0.003 0.04 ± 0.001 0.08 ± 0.003 0.10 ± 0.005 0.09 ± 0.004Arg 0.05 ± 0.001 0.20 ± 0.011 0.33 ± 0.018 0.30 ± 0.017 0.25 ± 0.012 0.30 ± 0.013 0.33 ± 0.016 0.30 ± 0.017 0.34 ± 0.018 0.21 ± 0.010 0.28 ± 0.010 0.35 ± 0.018 0.35 ± 0.018Val 0 0.21 ± 0.012 0.62 ± 0.030 0.42 ± 0.022 0.40 ± 0.022 0.42 ± 0.020 0.39 ± 0.022 0 0.04 ± 0.001 0.03 ± 0.001 0.42 ± 0.021 0.46 ± 0.022 0.44 ± 0.025Met 0.40 ± 0.018 0.03 ± 0.001 0.03 ± 0.001 0.03 ± 0.001 0.02 ± 0.001 0.02 ± 0.001 0.03 ± 0.001 0 0.03 ± 0.001 0.02 ± 0.001 0.02 ± 0.001 0.03 ± 0.001 0.04 ± 0.002Leu 0.21 ± 0.009 0.07 ± 0.002 0.22 ± 0.009 0.20 ± 0.009 0.17 ± 0.007 0.19 ± 0.007 0.09 ± 0.003 0.21 ± 0.011 0.09 ± 0.003 0.06 ± 0.002 0.06 ± 0.002 0.08 ± 0.003 0.22 ± 0.011Phe 0.02 ± 0.000 0 0 0.01 ± 0.001 0 0 0 0.02 ± 0.001 0.11 ± 0.006 0 0 0 0Ile 0 0.04 ± 0.001 0.04 ± 0.001 0 0.05 ± 0.002 0.06 ± 0.003 0.06 ± 0.003 0.05 ± 0.001 0.01 ± 0.001 0.01 ± 0.001 0.05 ± 0.002 0.06 ± 0.003 0.06 ± 0.003Tyr 0.03 ± 0.001 0 0 0.03 ± 0.001 0.03 ± 0.001 0 0 0.04 ± 0.002 0.04 ± 0.001 0.03 ± 0.001 0 0 0Trp 0.01 ± 0.000 0.08 ± 0.004 0.12 ± 0.006 0.11 ± 0.005 0.08 ± 0.004 0.94 ± 0.030 0.11 ± 0.005 0.10 ± 0.005 0.01 ± 0.000 0 0.09 ± 0.005 0.12 ± 0.007 0.12 ± 0.007Asp 0.04 ± 0.001 0.31 ± 0.011 0.72 ± 0.039 0.33 ± 0.017 0.50 ± 0.023 0.52 ± 0.025 0.58 ± 0.032 0.02 ± 0.001 0.08 ± 0.004 0.05 ± 0.002 0.54 ± 0.030 0.56 ± 0.020 0.54 ± 0.030Glu 0.01 ± 0.000 0.04 ± 0.002 0.05 ± 0.002 0.01 ± 0.000 0.01 ± 0.000 0.04 ± 0.002 0.05 ± 0.002 0.01 ± 0.000 0.04 ± 0.002 0.04 ± 0.002 0.04 ± 0.002 0.05 ± 0.002 0.05 ± 0.002Sum 10.62 ± 0.52 2.06 ± 0.08 3.06 ± 0.13 2.29 ± 0.11 3.08 ± 0.14 4.26 ± 0.17 25.73 ± 1.03 3.08 ± 0.11 24.55 ± 0.98 14.94 ± 0.56 2.99 ± 0.12 25.81 ± 0.95 3.53 ± 0.09Signe c a ab a ab b e ab e d ab e ab

a BCS103® yeast strain.b EC1118® yeast strain.c With Lallzyme MMX® addition.d γ-Aminobutyric acid.e Different letters indicate significant differences (p ≤ 0.01).

90S.Torresietal./Food

ResearchInternational55

(2014)83

–92

Fig. 3. Foam characteristics of the base wine (time 0) and sparkling wines produced withEC1118® andBCS103®yeast strains, in thepresence or in the absence of LallzymeMMX®,at 4, 8 and 12 months after tirage. Foam height (a); foam stability (b).

91S. Torresi et al. / Food Research International 55 (2014) 83–92

the observation time, it remained virtually constant and similar in allthe samples, independently from the yeast strain and the enzyme. Areduction in the values of foam height during wine aging with yeastwas observed by other authors (Maujean, Poinsaut, Dantan, Brissonet,& Cossiez, 1990).

Foam stability (Fig. 3b) showed a similar behavior. At 4 months, itremained constant (EC1118® sample without enzyme) or declinedgreatly (all other samples); subsequently (8 and 12 month aging), thevalues showed minor changes, with differences among wines at thesame time of sampling resulting not statistically significant (p ≤ 0.01).As observed for the foamheight, therefore, at the end of the observationtime, all wines manifested similar foam stability values, lower than attime zero (base wine).

We can note that, in the samples aged 12 months, both the heightand the stability of the foam were not influenced by either the strainor the enzyme addition, in spite BCS103®yeast showed a higher proteincontent with respect to EC1118®, at this aging time (Fig. 2). These find-ings were in contrast with those of Pueyo et al. (1995), who found thatprotein content of cavawines (sparklingwinesmade according to tradi-tional method) was negatively related with foam height and positivelywith foamstability. On theother hand, other authors did notfind relation-ships between protein concentration and foam characteristics (Maujeanet al., 1990). In any case, our results indicated that β-glucanase additionhad no influence either on total protein content or on foam characteristicsof the sparkling wines aged up to 12 months.

4. Conclusions

In this study, the application of LallzymeMMX® enzymatic prepara-tion to improve yeast autolysis during the aging of traditional sparklingwines was evaluated. We found that the β-glucanase contained in the

commercial enological preparation was slightly inhibited by ethanol,and that its residual activity at wine pH was sufficient for the purposes.Kinetic parameters obtained by the Michaelis–Menten kinetics showeda better enzyme–substrate complex formation for EC1118® strain, be-cause of its lower KM value.

The monitoring of on-lees aging process showed that cell lysis be-came evident 8 and 12 months after tirage for EC1118® and BCS103®,respectively, and that LallzymeMMX®was able to enhance the cell dis-organization for both yeasts, thus improving the release of amino acids(and probably other intracellular components) into the wine.

The results of our research demonstrated that the addition ofβ-glucanase might be a promising tool in accelerating the aging charac-teristics of traditional sparkling wines, but its effects varied accordingwith the strain used for secondary fermentation. In fact, enzyme influencewas more evident on the EC1118® strain, naturally provided with betterautolytic activity. Notwithstanding, β-glucanase employment might alsobe useful when, for various reasons, yeasts having minor natural lyticcapacity (like BCS103®) are utilized for refermentation.

We found that β-glucanase considerably influenced the free aminoacid profile of the sparklingwines, with yeast-dependent effects. There-fore, on considering that amino acids are involved in the formation of anumber of aroma substances, the impact of the enzyme on aroma com-pounds and sensorial characteristics of sparkling wines produced withdifferent yeast strains should be investigated. Moreover, further studiesare necessary to evaluate the influence of potential inhibitors (such aspolyphenols, bentonite and SO2) on enzyme activity.

References

Alexandre, H., & Guilloux Benatier, M. (2006). Yeast autolysis in sparkling wine — Areview. Australian Journal of Grape and Wine Research, 12, 119–127.

Bisson, L. F. (1991). Influence of nitrogen on yeast and fermentation of grapes. In J. M.Rantz (Ed.), Proceedings of the international symposium on nitrogen in grapes andwine (pp. 78–89). Washington, DC: The American Society for Enology and Viticulture.

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein–dye binding. AnalyticalBiochemistry, 72, 248–255.

Brissonnet, F., & Maujean, A. (1991). Identification of some foam active compounds inchampagne base wine. American Journal of Enology and Viticulture, 44, 97–102.

Charpentier, C., & Feuillat, M. (1993). Yeast autolysis. In G. H. Fleet (Ed.),Winemicrobiologyand biotechnology (pp. 225–242). London: Taylor and Francis.

DeMarco, J. L., & Felix, C. R. (2007). Purification and characterization of a β-glucanase pro-duced by Trichoderma harzianum showing biocontrol potential. Brazilian Archives ofBiology and Technology, 50, 21–29.

Dubourdieu, D., Villetaz, J. C., Desplanques, C., & Ribéreau Gayon, P. (1981). Dégradationenzymatique du glucane de Botrytis cinerea. Application à l'amélioration de la clarifi-cation des vins issus de raisins pourris. Connaissance de la Vigne et du Vin, 15, 161–177.

García,M. J., Aleixandre, J. L., Álvarez, I., & Lizama, V. (2009). Foamaptitude of Bobal varietyin white sparkling wine elaboration and study of volatile compounds. European FoodResearch and Technology, 229, 133–139.

Gonzalez, R., Martínez-Rodríguez, A. J., & Carrascosa, A. V. (2003). Yeast autolytic mutantspotentially useful for sparkling wine production. International Journal of FoodMicrobiology, 84, 21–26.

Humbert Goffard, A., Saucier, C., Moine Ledoux, V., Canal Llaubères, R. M., Dubourdieu, D.,& Glories, Y. (2004). An assay for glucanase activity in wine. Enzyme and MicrobialTechnology, 34, 537–543.

Luguera, C., Moreno-Arribas, M. V., Pueyo, E., Bartolomé, B., & Polo, M. C. (1998). Fraction-ation and partial characterization of protein fractions present at different stages ofthe production of sparkling wines. Food Chemistry, 63, 465–471.

Marchal, R., Seguin, V., & Maujean, A. (1997). Quantification of interferences in the directmeasurement of proteins in wines from the Champagne region using the BradfordMethod. American Journal of Enology and Viticulture, 48, 303–309.

Martínez-Rodríguez, A., Carrascosa, A. V., Barcenilla, J. M., Pozo Bayón, M.A., & Polo, M. C.(2001). Autolytic capacity and foam analysis as additional criteria for the selection ofyeast strains for sparkling wines production. Food Microbiology, 18, 183–191.

Martínez-Rodríguez, A. J., Carrascosa, A. V., Martín Àlvarez, P. J., Moreno-Arribas, M. V., &Polo, M. C. (2002). Influence of the yeast strain on the changes of the amino acids,peptides and proteins during sparkling wine production by the traditional method.Journal of Industrial Microbiology and Biotechnology, 29, 314–322.

Martínez-Rodríguez, A. J., Carrascosa, A. V., & Polo, M. C. (2001). Release of nitrogencompounds to the extracellular medium by three strains of Saccharomyces cerevisiaeduring induced autolysis in a model wine system. International Journal of FoodMicrobiology, 68, 155–160.

Martínez-Rodríguez, A. J., Polo, M. C., & Carrascosa, A. V. (2001). Structural and ultrastruc-tural changes in yeast cells during autolysis in a model wine system and in sparklingwines. International Journal of Food Microbiology, 71, 45–51.

92 S. Torresi et al. / Food Research International 55 (2014) 83–92

Masino, F., Montevecchi, G., Arfelli, G., & Antonelli, A. (2008). Evaluation of the combinedeffects of enzymatic treatment and aging on lees on the aroma of wine from Bombinobianco grapes. Journal of Agricultural and Food Chemistry, 56, 9495–9501.

Maujean, A. A., Poinsaut, P., Dantan, H., Brissonet, F., & Cossiez, E. (1990). Etude de latenue et de la qualité de mousse des vins effervescents. II. Mise au point d'une tech-nique de mesure de la moussabilité de la tenue et de la stabilité de la mousse desvines effervescents. Bulletin OIV, 63, 405–427.

McIlvaine, T. C. (1921). A buffer solution for colorimetric comparison. Journal of BiologicalChemistry, 49, 183–186.

Noronha, E. F., & Ulhoa, C. J. (2000). Characterization of a 29-kDA β-1,3-glucanase fromTrichoderma harzianum. FEMS Microbiology Letters, 183, 119–123.

Nuñez, Y. P., Carrascosa, A. V., González, R., Polo, M. C., & Martínez-Rodríguez, A. J. (2005).Effect of accelerated autolysis of yeast on the composition and foaming properties ofsparklingwines elaborated by a champenoisemethod. Journal of Agricultural and FoodChemistry, 53, 7232–7237.

Palomero, F., Benito, S., Morata, A., Tesfaye,W., González, M. C., & Suárez Lepe, J. A. (2009).Effect on the autolysis process and the colouring matter of several commercial prep-arations with β-glucanase action in red winemaking. European Food Research andTechnology, 229, 585–592.

Pérez-Serradilla, J. A., & Lauque De Castro, M.D. (2008). Role of lees in wine production: Areview. Food Chemistry, 111, 447–456.

Pozo Bayόn, M.Á., Martínez-Rodríguez, A. J., Pueyo, E., & Moreno-Arribas, M. V. (2009).Chemical and biochemical features involved in sparkling wine production: From atraditional to an improved winemaking technology. Trends in Food Science andTechnology, 20, 289–299.

Pretorius, I. S. (2000). Tailoring wine yeast for the new millennium: Novel approaches tothe ancient art of winemaking. Yeast, 16, 675–729.

Pueyo, E., Martín Álvarez, P. J., & Polo, M. C. (1995). Relationship between foam characteris-tics and chemical composition inwines andcavas (sparklingwines).American Journal ofEnology and Viticulture, 46, 518–524.

Rodriguez-Nogales, J. M., Fernández-Fernández, E., Gómez, M., & Vila-Crespo, J. (2012).Antioxidant properties of sparkling wines produced with β-glucanase and commer-cial yeast preparations. Journal of Food Science, 77, C1005–C1010.

Rodriguez-Nogales, J. M., Fernández-Fernández, E., & Vila-Crespo, J. (2012). Effect of theaddition of β-glucanase and commercial yeast preparations on the chemical and sen-sorial characteristics of traditional sparkling wine. European Food Research and Tech-nology, 235, 729–744.

Todd, B. E. N., Fleet, G. H., & Henscheke, P. A. (2000). Promotion of autolysis through theinteraction of killer and sensitive yeasts: Potential application in sparkling wine pro-duction. American Journal of Enology and Viticulture, 51, 65–72.

Torresi, S., Frangipane, M. T., & Anelli, G. (2011). Biotechnologies in sparkling wine pro-duction. Interesting approaches for quality improvement: A review. Food Chemistry,129, 1232–1241.

Van Rensburg, P., & Pretorius, I. S. (2000). Enzymes in winemaking: Harnessing naturalcatalysts for efficient biotransformations — A review. South African Journal of Enologyand Viticulture, 21, 52–73 (Special Issue).

Vázquez Garcidueñas, S., Leal Morales, C. A., & Herrera Estrella, A. (1998). Analysis of theβ-1,3-glucanolytic system of the biocontrol agent Trichoderma harzianum. Applied andEnvironmental Microbiology, 64, 1442–1446.

Wright, R. (2000). Transmission electron microscopy of yeast. Microscopy Research andTechnique, 51, 496–510.

Zinnai, A., Venturi, F., Quartacci, M. F., Andreotti, A., & Andrich, G. (2010). The kineticeffect of some wine components on the enzymatic hydrolysis of β-glucan. SouthAfrican Journal of Enology and Viticulture, 31, 9–13.