Abstract
Polysaccharides are important in colloidal stabilization-destabilization phenomena. In a previous study, three different trends were observed upon addition of wine polysaccharides to procyanidin suspensions: rhamnogalacturonan II monomers had no impact, the MP0 mannoprotein fraction strongly inhibited aggregation, and rhamnogalacturonan II dimers increased aggregation. After molecular weight fractionation of MP0 mannoprotein, the molecular weight effect on polyphenol particle size was investigated. Low molecular weight macromolecules are the stabilizing species, while high molecular weight molecules have no impact in “standard” wine conditions (pH 3.4 buffers containing 2 gL−1 tartaric acid and 12% ethanol). At high ionic strengths or low ethanol concentrations, all mannoprotein fractions prevented tannin aggregation and precipitation.
Polyphenolic compounds are widely present in many beverages and important for both stability and sensory properties. Native condensed tannins found in wines are flavan-3-ol oligomers and polymers. Grape seed procyanidin fractions form colloidal particles in hydroalcoholic solutions (Poncet-Legrand et al. 2003, Riou et al. 2002, Saucier et al. 1997). Tannins are also known for their ability to complex with proteins (Luck et al. 1994, McManus et al. 1985, Waters et al. 1994b), which is important both technologically (gelatin addition during fining) (Maury et al. 2001) and sensorially (astringency) (Maury et al. 2001, Sarni-Manchado et al. 1999). Several studies have been conducted on tannin-protein interactions, mainly by nuclear magnetic resonance (NMR) (Charlton et al. 2002a,b, Richard et al. 2001), microcalorimetry (Frazier et al. 2003), turbidimetry (de Freitas and Mateus 2001, Siebert et al. 1996, Siebert and Lynn 1997), and more recently dynamic light scattering (Poncet-Legrand et al. 2006). However, less research has been done on polysaccharide-polyphenol interactions. Polysaccharides, which are naturally present in wine, can interact with tannins and modify polyphenol/polyphenol and polyphenol/protein interactions. Adding polysaccharides to a model tannin suspension has a strong impact on procyanidin aggregation (Riou et al. 2002). For example, addition of a small amount of rhamnogalacturonan II (RGII) dimer increased polyphenol aggregation, at times leading to precipitation, but addition of acid arabinobinogalactan AGP4 and mannoprotein MP0 limited their aggregation and prevented their precipitation.
Wine mannoproteins play an important part in protein haze stabilization (Dupin et al. 2000a,b, Moine-Ledoux and Dubourdieu 1999, Waters et al. 1994a). They have been identified as either high molecular weight (~420 kDa) mannoproteins (Waters et al. 1994a) or low molecular weight (~32 kDa) polysaccharides (Moine-Ledoux and Dubourdieu 1999). In order to clarify the stabilization mechanisms, we studied the molecular weight effect, as polymers adsorbed on particle surfaces play a decisive part in colloidal stabilization (Evans and Wennerström 1999). For example, high molecular weight linear polymers induce flocculation, whereas lower molecular weight linear polymers act as steric stabilizers. To assess this effect in our systems, the MP0 mannoprotein fraction was further fractionated into three subfractions, differing mainly by molecular weight (Vidal et al. 2003). Here we present results obtained when these MP0 fractions were added to a polyphenol colloidal suspension in model wine buffer (typically aqueous solutions containing 12% ethanol, pH adjusted to 3.4 with tartaric acid and sodium hydroxide). Morover, as ethanol concentration and ionic strength affect polyphenol aggregation (Poncet-Legrand et al. 2003), we performed the same experiments with mannoproteins (ethanol concentration ranging from 5% to 20%; ionic strength ranging from 10−3 to 10−1 M) to describe the driving forces involved in aggregation and/or stabilization.
Materials and Methods
Chemicals.
Deionized water was obtained with a Milli-Q system (Millipore, Billerica, MA). Absolute ethanol, acetone, and glycerol (all analytical grade) were purchased from Prolabo (Fontenay sous Bois, France), tartaric acid and sodium chloride from Labosi (Illkirch, France), and sodium hydroxide (analytical grade) from Merck (Darmstadt, Germany).
Purification and characterization of tannin fractions.
Grape seed tannins, which are flavan-3-ol polymers, were purified from freeze-dried seeds of Vitis vinifera, as described by Poncet-Legrand et al. (2003). The tannin fraction studied here was the most abundant, corresponding to the F3 fraction described previously (Poncet-Legrand et al. 2003), and was analyzed by high-performance liquid chromatography (HPLC) after thiolysis (Prieur et al. 1994, Souquet et al. 1996) to estimate the mean degree of polymerization (mDP) as 11 and 27% galloylation.
Polysaccharides.
The MP0 mannoprotein fraction used in the present study was purified from a red wine as described previously (Pellerin et al. 1996, 1995, Vernhet et al. 1996). The MP0 mannoprotein was obtained after two anion-exchange chromatography steps: it was twice the unretained fraction after elution in citrate (pH 4.6) then acetate (pH 5.4) buffers, and its charge was low in comparison with the MP1, MP2, and MP3 fractions (Vernhet et al. 1996). The MP0 fraction was then separated by size-exclusion chromatography (Vidal et al. 2003) into three subfractions a, b, and c (Table 1⇓). Therefore we assumed that the main difference was the molecular weight.
Sample preparation.
Polysaccharide stock solutions were prepared in water and stored at −40°C. Stock solutions of tannins in absolute ethanol were prepared, degassed, and stored at −40°C. These were used within the week to avoid tannin oxidation. Ethanol was chosen to ensure a complete dissolution of tannins. Tannin and polysaccharide stock solutions were diluted in buffer solution (2 g L−1 tartaric acid, unless stated otherwise). The pH was adjusted to 3.4 with 5 M sodium hydroxide, and aggregation kinetics were monitored by dynamic light scattering as described previously (Poncet-Legrand et al. 2003). Ethanol concentration and ionic strength of the model solutions were varied. The effect of ethanol concentration (5, 12, 20%) on aggregation was studied at a constant ionic strength of 10−2 M, while the effect of ionic strength (10−3, 10−2, 10−1 M) was investigated at 12% ethanol. The composition of the different buffers is given in Table 2⇓.
Dynamic light scattering.
Dynamic light scattering was measured with an Autosizer 4800 (Malvern Instruments, Malvern, UK) equipped with a 35 mW He-Ne laser (Coherent, Auburn, CA) and at a wavelength of 633 nm. The sample cell was thermostated at 25 ± 0.1°C. Measurements were carried out at a 90° angle from the incident beam. The time-dependence of the light scattering was monitored and the autocorrelation function G(t) of the particles was measured. The diffusion coefficient of the particle (D) was derived from this function, and hydrodynamic radii of the particles RH were calculated using the Stokes-Einstein equation. All calculations were made assuming that particles were spherical and thus this equation can be written as:
where kB is the Boltzmann constant, T is temperature, and η is solvent viscosity. The cumulant method was used to fit the autocorrelation curves and the results were referred to as Zav (average size), which is the mean hydrodynamic diameter. Time zero was set at the mixing of stock solutions with buffer, and each measurement was the average of 10 sub-runs.
Viscosity measurements.
Viscosities of the water-ethanol mixtures were measured with an AVS 400 capillary viscosimeter (Schott Geräte GMBH, Hofheim, Germany) in a CT 050 bath thermostated at 25°C (Schott Geräte). Measurements were done in triplicate (Table 2⇑).
Results and Discussion
Molecular weight effect (MP0a, b, c).
When stock solutions of grape seed tannins were diluted in tartaric acid buffer, colloidal particles were formed, the size of which increased with time (Figure 1⇓). The kinetics of tannin aggregation were followed by dynamic light scattering. Without polysaccharide, particle size increased rapidly, and after 12 hr there was still no equilibrium. When 0.05 g L−1 MP0c was added, no effect was observed. For concentrations of 0.1 g L−1 and above, aggregation was slowed and/or stopped (Figure 1⇓). The same experiments were performed with the two other mannoprotein fractions, in the same range of concentrations, and particle average diameters were plotted after 4 hr (Figure 2⇓). Two different outcomes were observed: the high molecular weight MP0a had almost no effect on particle size, whereas the lower molecular weight fractions MP0b and MP0c stabilized polyphenol aggregates at concentrations of 0.1 g L−1 and above.
Polymers can act as stabilizers or flocculating polymers, depending on their molecular weight, charge, and structure. High molecular weight cationic and neutral polymers are used in water treatment to flocculate colloidal aggregates, while low molecular weight polymers are used to stabilize latexes and inorganic particles. Flocculation is usually explained by bridging: the same macromolecule can bind to several particles, allowing the formation of bigger aggregates. In our experiment, the trend was not that acute, but we were monitoring a dynamic process and aggregation occurred after a few hours. The mannoproteins we used were linear polymers, with a very high degree of branching (presence of 2,6 linked mannose in the MP structure). At low mannoprotein concentrations (0.05 g L−1), the polyphenol particle surface coverage is not high enough for steric stabilization to occur. At higher polysaccharide concentrations (higher surface coverage), lower molecular weight mannoproteins provide good conditions for steric stabilization (Figure 3⇓).
Ethanol concentration effect.
Ethanol concentration strongly affects tannin aggregation. We studied the effect of adding 0.15 g L−1 polysaccharides to a tannin solution (1 g L−1) as ethanol increased from 5 to 20% (Figure 4a⇓). The 0.15 g L−1 polysaccharide concentration was chosen because at this concentration in 12% ethanol, MP0b began to have a strong effect on tannin aggregation. In 20% ethanol, polyphenols formed relatively stable particles (Poncet-Legrand et al. 2003), but at 5% ethanol there was strong tannin aggregation leading to visual phase separation (precipitation, particle size >1 μm) after several hours (Figure 4b⇓).
All low and high molecular weight mannoprotein fractions prevented strong aggregation of polyphenols. At 20% ethanol, there was no polyphenol precipitation (Poncet-Legrand et al. 2003). The particles formed were stable, the polyphenols were in good solvent (in Flory-Huggins theory, the solvent gives rise to an exothermic enthalpy of mixing), and thus polysaccharides did not have a striking effect. At 5% ethanol, the tannins were in a poor solvent (Poncet-Legrand et al. 2003) (enthalpy of mixing was positive), and after a few hours tannin precipitation was observed if no polysaccharide was added. In contrast, polysaccharides were in good solvent (more water, less ethanol) and provided better steric stabilization.
Ionic strength effect.
The final experiments involved varying ionic strength (I) from 10−3 to 10−1 M. A previous study on polyphenols alone showed that adding salt increased polyphenol/polyphenol interaction by increasing the hydrophobic interactions (Poncet-Legrand et al. 2003). Reference experiments were performed on mannoprotein solutions at 0.15 g L−1 in all buffers: ethanol 12% and tartaric acid 0.2 and 2 g L−1, 2 g L−1 + 0.1 M NaCl, as well as tartaric acid 2 g L−1 and ethanol 5 and 20%. In all cases, no particle was detected: the scattered intensity was very close to that of the buffer, no correlation function could be measured, and the signal/noise ratio was poor. From a macroscopic point of view, formation of polysaccharide aggregates was not favored by addition of salt. That is not surprising, as MP0s, unlike the MP1, MP2, and MP3 mannoprotein fractions, have little or no electric charge (Vernhet et al. 1996): adding salt does not screen electrostatic repulsions, thus no aggregation or further stabilization is expected.
At low salt concentration, tannin aggregation occurred very slowly and particle size remained low. The polysaccharide effect was thus not striking (Figure 5⇓). When ionic strength was increased, polyphenol aggregation also increased. Adding mannoproteins of any molecular weight protected polyphenol aggregates from precipitation. We attribute this stabilization to steric stabilization by the mannoproteins. At high ionic strengths, even high molecular weight polysaccharide was an efficient polyphenol particle stabilizer.
Conclusion
A previous study showed that some polysaccharides act as particle stabilizers. Present results confirm that the mechanism involved is probably steric stabilization: medium and low molecular weight polymers are more efficient than high molecular weight polymers, and all polysaccharides tested prevented polyphenol precipitation when ionic strength was increased. That allowed us to dismiss an electrostatic stabilization mechanism: if such a mechanism was involved, adding salt would screen electrostatic repulsions and thus favor aggregation. The results reported here must be confirmed with other polysaccharide fractions, with complementary methods such as microcalorimetry and NMR, and on other polyphenol fractions such as skin tannins. Further studies should also be conducted on the mannoprotein effect with well-dissolved tannins rather than aggregates.
- Received April 2006.
- Revision received July 2006.
- Copyright © 2007 by the American Society for Enology and Viticulture