ABSTRACT

Unstable proteins can form a haze in juice or wine. Bentonite fining is the method used by processors to remove them. However, Bentonite generates waste that must be managed.

To determine whether resins could replace the Bentonite fining practice, three commercial adsorbent resins were tested for their protein adsorbent properties. The resins were tested using a pumped-flow pilot scale column system. Resin-treated juices and wine were analyzed for protein stability, the quantity of protein and phenolics removed, titratable acidity, and metals to quantitate any processing effects.

The resins effectively stabilized the grape juices and wine. Removal of protein and phenolics levels in untreated grape juices and wine were followed up to ~ 100 Bed Volumes through each of the columns. Resin treatment did not affect the pH, titratable acidity, or common cation content of the juices or wine. The sugar content of the grape juice was also unchanged. The resins also removed some of the heavy metals. Some color and aroma were also removed.

INTRODUCTION

Red wine usually has a high polyphenol content and materials such as tannins complex with proteins and precipitate from the wine at an early stage of wine-making; therefore, there is less problem of protein instability after cellar operations (8).

White wine and wines of low polyphenol content, however, pose special concerns regarding the potential for haze formation after bottling. The color and clarity of a wine are important aspects affecting one’s decision of whether or not to purchase a bottle of white wine.

The differences in proteins due to climate, cultivar, grape maturity, electrical charge, and interaction with other components make the nature of protein instability difficult to explain. The major source of proteins in wine is the grape. Autolysis of yeast cells following fermentation is a secondary source of protein in wine (8).

Fining before bottling is used to avoid haze problems (8). Fining is "the removal or modification of certain undesirable components from a wine in order to improve its appearance and/or organoleptic qualities" (1). Unstable protein can be removed by a variety of methods such as bentonite adsorption, silica gel adsorption, tannic acid fining, proteolytic enzyme treatment, and ultrafiltration (3). Of the above, bentonite fining is a common practice in wineries today. The mechanism of fining includes coagulation and mechanical entrapment. Coagulation works because colloidal particles bind with oppositely charged fining agents. They cluster together and increase in density, and eventually precipitate out of solution. Mechanical entrapment is a co-mechanism of coagulation. While a cluster settles, the complex becomes a net that traps other impurities along the way. Calcium, magnesium, and potassium salts in the wine enhance the function of fining as well (8).

However, the protein fragments responsible for haze formation are not those primarily removed by either bentonite fining or ultrafiltration. The mechanisms of bentonite fining work on the removal of proteins that have sizes ranging mainly from 32000 to 45000 Dalton and isoelectric points (PI) between 5.8 and 8.0; those protein fragments represent only a small portion of soluble proteins. The proteins and glycoproteins that cause haze have sizes ranging from 12600 to 30000 Daltons and have PIs between 4.1 and 5.8 (1).

Studies have showed that haze formation in beverages relates to the concentrations of haze-active protein and haze-active polyphenol, and to their ratio. A model made of gelatin and tannic acid suggests that the fixed numbers of binding sites on proteins will bind to a fixed number of polyphenols by hydrogen and hydrophobic bonding. The amount of protein alone does not predict haze formation. Data indicate that the total protein analysis does not correspond to protein instability. Size rather than total amount of protein is more important in protein instability (4).

As mentioned earlier, protein stability related to clarity at the post-bottling stage of production is important for marketability. However, consumers will taste the wine afterwards, so the palate is important as well. An excessive amount of bentonite will strip out flavor or add an earthy smell; therefore, wineries have to compromise between clarity and sensory aspects. Bentonite fining also produces large volumes of lees, which lead to loss of wine and waste materials to be managed. Wineries can distill the alcohol content out of lees, but waste still needs to be handled. In the past, this solid waste would go back into vineyard, as fertilizer. Now anything containing alcohol is considered a hazardous waste and needs to be specially disposed. There are significant costs involved in handling de-alcoholized solid wastes, as well. Finding alternative ways of removing unstable protein in wine therefore is of great value.

Adsorbent resins have been used in the food industry to decolorize sugar juices, stabilize beverages against haze formation, and debitter citrus juice (2). They are promising materials for replacing bentonite in the fining process.

The mechanisms for removal of unstable proteins and polyphenols by resins include adsorption, size exclusion, and ion exchange (7). Organophylic adsorption occurs when two non-polar molecules are held together by Van der Waals forces. The benzene rings on the synthetic resin and phenolics enhance this particular adsorption. Phenolics that are fixed onto the resin matrix can bind with the proteins from juices or wines retaining proteins on the resin. To select a synthetic adsorbent, one must look at its adsorption capacity, as well as its effect on other components in grape juice and wine.

Three different resins, Rohm and Haas Amberlite XAD-16, Dow chemical XUS-40285, and Dow Chemical XU-43520 were used to study protein removal from white grape juice and wine. Amberlite XAD-16 is a styrene/divinylbenzene based material that is a nonionic, hydrophobic, crosslinked polymer with a surface having an aromatic nature. The adsorptive properties come from its macroreticular (macroporous) structure, which contains both a continuous polymer phase and a continuous porous phase. Pore size ranges from 2 to 300 Angstroms; the surface area is 800 m2/g. The distribution of pore size shows its application for adsorption of low to medium molecular weight organic compounds, especially proteins (Rohm and Haas company, product literature). The active sites on the resin adsorb protein; also the pores allow protein molecules to be trapped and separated from the sample. Polyphenols will be adsorbed by XAD16 resin by Van der Waals forces between two aromatic rings (6).

Two developmental polymeric adsorbents studied were Dow Chemical XUS-40285 and XUS-43520. The former is made of dimethylamine functionalized chloro-methylated copolymers of styrene and divinylbenzene. This resin has high surface area and high porosity. It is a weak base anion exchange resin. The protein removal mechanisms include ion exchange, size exclusion, and possible adsorption. Dow Chemical XU-43520 is a macroporous adsorbent resin, a styrene-divinylbenzene polymer somewhat similar to XAD-16 (Dow chemical company, material safety data sheet, 1997).

For each of these three resins, their protein adsorbent abilities were evaluated in terms of how much juice or wine could pass through a resin column as protein stable. Heat tests were used as the standard for determining protein haze stability. Samples, heated in an oven for a fixed period of time, were observed with the naked eye and a turbidimeter.

The amount of protein existing in a resin treated juice or wine was estimated chemically by the Bradford method. Coomassie Brilliant Blue (CBB), a dye, is able to fix proteins. Although both alcohol and endogenous grape berry phenolics can interfere with the analysis and lead to higher absorbance readings, the Bradford method is used for its reproducibility and rapidity (4).

MATERIALS AND METHODS

Adsorbent Resins:
(A) Amberlite® XAD16 (Rohm & Haas)
This resin is a nonionic, hydrophobic, crosslinked polymeric adsorbent. Its average pore size distribution (average pore diameter = 144A) makes this resin an excellent choice for the adsorption of proteins and high molicular weight colorants. It can be regenerated/eluted with water where adsorption is from an ionic solution such as with juices and wines.

(B) XUS40285 (Dow Chemical)
This resin is a hydrophilic adsorbent that demonstrates good wettability and compatibility with acid and base regenerating agents. High surface area and high porosity characterize this adsorbent. It also has been de-signed to increase bulk movement in and out of the bead via macropores. This resin mimics activated carbon as it has a large number of small pores or adsorption sites.

(C) XU43520 (Dow Chemical)
This macroporus adsorbent resin yields results similar to that of the XAD16 and does not require wetting in ethanol. However this resin does require stronger base and acid conditioning solutions.

Samples:
White grape juice
White wine fermented from Thompson Seedless grapes

Columns:
2.54 cm x 183 cm Pumped at the rate of ~ 7 Bed Volumes (BV)/hr using a Masterflex® peristaltic pump. 1-100 RPM. CAT. No. 755-30. with Masterflex® L/S three-roller pump heads. Model 7017-20 (Cole Parmer Instrument Co., Chicago, Illinois).

Filtered grape juice or wine was pumped into each of three columns with samples of processed juice and wine collected at approximate five BV intervals. These samples were subsequently analyzed for several physical and chemical parameters:

Heat stability/clarity:
Samples held at 49°C for 24 hours – cool and measure clarity visually and in nephelometric turbidity units (NTU) – Turbidimeter DRT-200B, Shaban Mfg., Inc.

Chemical protein levels:
Sigma® Microprotein. CAT. No. 610-A. Includes protein dye reagent: Brilliant Blue G, 35 mg/dL, in phosphoric acid and methanol. Protein standard solution: Human albumin, 30 mg/dL, in saline with sodium azide, 0.1%, as preservative.

Phenolics:
Folin-Ciocalteu method – The equipment and reagents employed in this analysis were identical, in most respect, to those suggested by Zoecklein et al. (8), except the reagents for determination of non-flavonoids: the Formaldehyde solution 0.8% was made from mixing 2.08 mL of 37.2% formaldehyde into 100 mL of deionized water, and the 1:5 hydrochloric acid was made by mixing 50 mL of concentrated hydrochloric acid (36.5-38%) into 200 mL of deionized water.

Misc. measurements:
pH, °Brix (wt% sucrose), titratable acidity (NaOH titration), metals (Na, K, Cu, Fe). The materials and procedures used are listed in Zoecklein et al. (8).

Operating conditions:
The operating conditions, including bed volume and flow rate for a typical resin run are those presented in Table 1.

RESULTS AND DISCUSSION

The results from this investigation of the abilities of adsorbent resins to remove proteins and phenolics from white grape juice and wine can be seen in Table 2.

The control juice (without resin treatment) had a measured turbidity of 23.8 NTU after being heated in the 49½C oven for 24 hours. When observed with the naked eye, there were both suspension and sediment present. The resin treated juice samples started to show haziness after heat treatment at 80BV for both column A (8.5NTU) and column C (7.2NTU). This occurred at 60BV for column C (11.6NTU). This means that both column A and column C have better abilities than column B in stabilizing Thompson seedless grape juice.

With the wine samples results are based solely on observations with the naked eye immediately following the heat treatment. Whatever haze was observed disappeared when these samples cooled down. For that reason, turbidity readings were not collected. Column A showed haze formation at 100BV, column B at 60BV, and column C at 110BV. This again indicates both column A and C have better abilities for stabilizing protein than column B.

Graphs of chemical protein level (Fig. 1 & Fig. 4) and total phenolics (Fig. 2 & Fig. 5) indicate significant early BV removals of these compounds with gradual increases in their concentrations, as the numbers of BV increased. Flavonoid phenol removals from juice (Fig. 3) follow the same pattern as total phenolics. The resin demonstrating the greatest removal and capacity for protein was Resin C, the Dow XU43520. The least efficient resin for protein removal, Resin B, Dow XUS40285, was more efficient in removing flavonoid and non-flavonoid phenols.

There were indications from color measurements on undiluted resin-treated juices and wines (measured at 420 nm – see Zoecklein, et al. [8]) that some color was removed. Typical results are presented in Table 3.

While resin treatment did not affect the potassium and sodium levels, there appeared to be some removal of copper and iron from juice (Table 4) [copper and iron levels were non-detect in the control wine, precluding any evaluation of removal]. The three resins appeared to be equally effective in removing copper. Resin B was more efficient in lowering iron in the treated juice (Fe 55% removal with Resin B vs. 22% for Resin A, and no removal with Resin C). This may indicate that these metals – especially copper – associate with phenols in the juice. Protein, phenolics, and metal removals occurred without significant changes in the control sample levels of pH, titratable acidity, and Brix.

The resin-treated wine was subjected to a preference test. The wine samples were scored so the higher the percentage the more favorable the wine was to a taste panel member. From Table 5, it can be seen that the panel noted the control wine without any treatment as the most favored.

CONCLUSIONS

Three commercially available adsorbent resins have demonstrated abilities to remove proteins and phenolic compounds responsible for formation of "protein haze" in white grape juice during pilot scale pumped-flow runs.

The resin demonstrating the greatest removal and capacity for protein was Resin C, the Dow XU43520. The least efficient resin for protein removal, Resin B, Dow XUS40285, was more efficient in removing flavonoid and non-flavonoid phenols. This latter resin was also more efficient in lowering copper and iron in the treated juice. None of the resins removed any significant amounts of potassium or sodium. The protein, phenolics, and metal removals occurred on all three resins without significant changes in the control juice and wine levels of potassium, pH, titratable acidity, and Brix. Total protein alone does not predict well the haze formation in the juice or wine samples. The phenolic compounds also played an important role.

The resin treatment produced colorless grape juice in early bed volumes. The resin also removed some of the phenolics, so the juice was less astringent (as noted by the sensory panel). These two characteristics can be valuable in a product used as a sweetener for other beverages. As for wine, since the early bed volumes of resin treated wine are extremely protein stable, it is possible that they can be blended back into untreated wine and still prevent haze formation. The fact that the sensory panel members preferred wine treated by Dow XUS-40285 (B) resin next to the control wine shows promise for resin applications in the juice and wine industries.

For future work, the means of regeneration, the optimal flow rate, and the cost of the operation are to be examined. More sensory evaluation on the blended juice or wine products will also be carried out.

ACKNOWLEDGEMENTS

The authors would like to thank the following for their support of this research: The California Agricultural Technology Institute, Canandaigua Wine Company, the American Vineyard Foundation, Viticulture and Enology Research Center, and California State University, Fresno Department of Chemistry.

LITERATURE CITED

1. Hsu, J. C., and D. A. Heatherbell. Heat-unstable proteins in wine. I. characterization and removal by bentonite fining and heat treatment. Am.J. Enol. Vitic. 38:11-15 (1987).
2. LaFlamme, J., and R. Weinand. New developments by the combination of membrane filtration and adsorption technology. Fleussiges obst. 60:336-342 (1993).
3. Marchal, R., V. Seguin, and A. Maujean. Quantification of interferences in the direct measurement of proteins in wine from the champagne region using the Bradford method. Am. J. Enol. Vitic. 48:303-309 (1997).
4. Siebert, K. J., N. V. Troukhanova, and P. Y. Lynn. Nature of polyphenol-protein interactions. J. Agric. Food Chem. 44:80-85 (1996).
5. Siebert, K. J., A. Carrasco, and P. Y. Lynn. Formation of protein-polyphenol haze in beverages. J. Agric. Food Chem. 44:1997-2005 (1996).
6. Weinand, R. Adsorbent resins in the beverage industry. Fruit Processing 5:166-171 (1995).
7. Willard, H. H., Lynne L. M., Jr., J. A. Dean, and F. A. Settle, Jr. Instrumental Methods of Analysis (7th ed). Wadsworth Publishing Company, Belmont, California (1988).
8. Zoecklein, B. W., K. C. Fugelsang, B. H. Gump, and F. S. Nury. Wine Analysis and Production. Chapman & Hall, New York, New York (1995).

 

---

{ CATI , VERC , VERC - Research Publication }

 

Copyright © 1999. All rights reserved.
CALIFORNIA AGRICULTURAL TECHNOLOGY INSTITUTE - CATI
College of Agricultural Sciences and Technology
California State University, Fresno