Food Chemistry 160 (2014) 16–21
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Impact of condensed tannin size as individual and mixed polymers on bovine serum albumin precipitation James F. Harbertson a, Rachel L. Kilmister b, Mark A. Kelm c, Mark O. Downey b,⇑ a
School of Food Science, Washington State University, Irrigated Agriculture Research and Extension Center, 24106 N. Bunn Rd., Prosser, WA 99350-8694, USA Department of Environment and Primary Industries, Victoria, PO Box 905, Mildura, Vic., Australia c Constellation Brands Inc., Mission Bell Winery, Research and Development, 12667 Road 24, Madera, CA 93637, USA b
a r t i c l e
i n f o
Article history: Received 18 October 2013 Received in revised form 4 March 2014 Accepted 8 March 2014 Available online 15 March 2014 Keywords: Condensed tannin Proanthocyanidin Theobroma cacao Protein precipitation Bovine serum albumin (BSA) Astringency
a b s t r a c t Condensed tannins composed of epicatechin from monomer to octamer were isolated from cacao (Theobroma cacao, L.) seeds and added to bovine serum albumin (BSA) individually and combined as mixtures. When added to excess BSA the amount of tannin precipitated increased with tannin size. The amount of tannin required to precipitate BSA varied among the polymers with the trimer requiring the most to precipitate BSA (1000 lg) and octamer the least (50 lg). The efficacy of condensed tannins for protein precipitation increased with increased degree of polymerisation (or size) from trimers to octamers (monomers and dimers did not precipitate BSA), while mixtures of two sizes primarily had an additive effect. This study demonstrates that astringent perception is likely to increase with increasing polymer size. Further research to expand our understanding of astringent perception and its correlation with protein precipitation would benefit from sensory analysis of condensed tannins across a range of polymer sizes. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Tannins are plant secondary metabolites that play a role in plant defence; mainly as a deterrent to herbivory (Hagerman & Butler, 1991; Harborne & Grayer, 1993), although there is some evidence of defence against microorganisms (Dixon & Lamb, 1990). Tannins also play an important role in a number of economically important areas of agricultural production; prevention of pasture bloat (Li, Tanner, & Larkin, 1996), the tanning of hides (Haslam, 1998) and in the mouthfeel of red wines (Gawel, 1998; Malien-Aubert, Dangles, & Amiot, 2002; Ribéreau-Gayon, 1982). In most of these systems, it is the ability of tannin to precipitate protein that underpins their function (Haslam, 1998). As a result the interaction of protein with tannin has attracted considerable research effort. Tannins are broadly grouped into two classes; hydrolysable and condensed tannins, both derived from plants. Hydrolysable tannins, which occur widely in wood, are based on gallic acid and occur as aggregations of gallates, known as ellagitannins, or are comprised of gallic acid moieties bound to a central glucose, e.g. penatgalloyl glucose and known as gallotannins (Haslam, 1998). Condensed tannins are polymers of flavanol subunits and are common in soft plant tissues such as leaves and fruit (skin and seeds)
⇑ Corresponding author. Tel.: +61 3 5051 4565; fax: +61 3 5051 4523. E-mail address: [email protected] (M.O. Downey). http://dx.doi.org/10.1016/j.foodchem.2014.03.026 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.
(Haslam, 1998). Tannins are a heterogeneous mixture containing a range of different polymer sizes, subunit compositions, and subunit linkages (Haslam, 1998). In grape seed alone it has been estimated that there is in excess of 65532 unique chemical tannin structures (Adams & Harbertson, 1999). Studies of tannin-protein interactions to date have given a strong indication that protein precipitation increases with increasing size (Porter & Woodruffe, 1984) and that the mode of interaction between tannins and proteins is a combination of hydrogen bonding and hydrophobic interactions (Charlton et al., 2002; Hagerman, Rice, & Ritchard, 1998). However, a limitation of all of the previous work was that owing to the complexity of isolating and purifying individual tannin polymers, these experiments have been conducted on mixtures of tannins with a range of polymer sizes and variable subunit composition (Porter & Woodruffe, 1984). The risk in using ill-defined mixtures and average polymer length to define tannin interactions and their correlation with astringency is the assumption that all tannins within a mixture interact with protein similarly. There have been no studies to date that document the interaction of thoroughly characterised individual condensed tannin polymers with protein. Consequently, several essential aspects about how tannins interact with protein are not well understood. There have been no studies of tannin-protein interactions that directly compare the protein precipitation of individual condensed tannins of increasing polymer size to unequivocally demonstrate the
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
relationship between polymer size and precipitation. The minimum molecular weight tannin required for protein precipitation has not yet been reported although it is thought that tannins less than tetramer do not precipitate bovine serum albumin protein (Adams & Harbertson, 1999). The impact on protein precipitation of structural features such as subunit composition is unclear as is the effect of different linkages (i.e. branched versus linear polymers). To address these questions, a methodology that would allow isolation of individual polymers was required. Such a method was developed for the separation of cacao polymers (Kelm, Johnson, Robbins, Hammerstone, & Schmitz, 2006). Condensed tannins from cacao are comprised of a single flavanol subunit, epicatechin, and can be readily separated up to octamer, although yields decrease with increasing polymer size. Using this method it has been possible to isolate sufficient quantity of pure polymers to determine the efficacy of tannin of different sizes for precipitating protein and the synergistic effect of these tannins in mixtures. Our hypotheses were that the efficacy of tannin for precipitating protein increases with size and the efficacy of polymer mixtures for precipitating protein is dependent upon polymer size.
17
liquid chromatography (HPLC) system using a Develosil diol column (300 50 mm i.d. 10 lm) in the hydrophilic interaction liquid chromatography (HILIC) mode. The binary mobile phase consisted of (A) acetonitrile and (B) methanol/water/formic acid (99:0.95:0.05, v/v/v). Gradient conditions for separation and collection were 0–30% solvent B from 0 to 35 min, 30% isocratic solvent B from 35 to 65 min, 30–80% solvent B from 65 to 67 min, and 80% isocratic solvent B from 67 to 80 min. Prior to injection the column was allowed to equilibrate (0% solvent B, 10 min). Cacao tannin (1 g) was dissolved in a 6 mL mixture of 25:75 mobile phase A/B, centrifuged (16,100g, 10 min), and filtered through a 0.45 lm PTFE filter prior to injection (5000 lL). The flow rate was set at 55 mL/min and the column temperature was maintained at room temperature (22 °C). Elution of separated cacao tannins was monitored by diode array detection (DAD) at 280 nm. Separated peaks were collected as they eluted, dried by rotary evaporation under reduced pressure (40 °C) then freeze dried to remove residual water and formic acid. Cacao fractions were prepared by this method three times for analysis. 2.4. HPLC-ESI-MS analysis of separated cacao tannins
2. Materials and methods 2.1. Chemicals Raw unfermented organic cacao (Theobroma cacao, L.) seeds of Equadorian origin were obtained from Natural Zing (Mt. Airy, MD, USA). Acetonitrile, methanol, hexane and ethyl acetate solvents were purchased from Merck (Darmstadt, Germany). Bovine serum albumin, sodium dodecyl sulphate, sodium hydroxide, triethanolamine, hydrochloric acid, acetic acid, formic acid, ferric chloride, ammonium formate, Amberlite resins, phosphorous pentoxide, ( )-epicatechin and (+)-catechin were purchased from Sigma-Aldrich Pty. Ltd (Castle Hill, Australia). HPLC columns were purchased from Phenomenex (Torrance, CA, USA). 2.2. Cacao tannin extraction and isolation Cacao seeds (500 g) were frozen with liquid nitrogen and ground in an electric coffee grinder to a 325 mesh (44 microns particle size). Ground cacao was defatted with hexane in a Soxhlet extractor for 6 hours. Defatted cacao was air dried overnight in a fume hood. To extract cacao tannin, a 100 g sample of defatted cacao was extracted twice with 1 L of 70% aqueous acetone on an orbital mixer (250 rpm; 30 min). To remove solids, the sample was filtered under vacuum through a glass microfiber filter. The filtrate was rotary evaporated under reduced pressure (40 °C) to remove acetone. The sample was mixed twice with 500 mL of ethyl acetate to remove low molecular weight material (epicatechin and tannin dimers). The aqueous phase was retained and mixed with Amberlite FPX62 resin (500 mg/mL) using an orbital mixer (250 rpm; 8 h). The sample was filtered under vacuum through a glass microfiber filter to remove the resin and the filtrate was then loaded on a glass column packed with Amberlite XAD7HP (20–60 mesh, 500 cm3) preconditioned with water (2 L). The column was washed with water (10 L) to remove any remaining anthocyanin, sugars and organic acids, then eluted with 80% aqueous ethanol (2 L). The eluate was rotary evaporated under reduced pressure (40 °C) to near dryness and further dried under vacuum with phosphorous pentoxide to yield a red-purple crusty solid (8.1 g). 2.3. Preparative HPLC fractionation of cacao tannin The fractionation of cacao tannin by degree of polymerisation was achieved on an Agilent 1100 preparative high performance
The qualitative analyses of separated cacao tannins was performed by HPLC (Agilent 1100) interfaced with an ion trap mass spectrometer (G2445D Bruker, Billerica, MA, USA). HPLC analysis was performed on a 250 4.6 mm i.d. 5 lm Develosil diol column fitted with a guard column of the same phase. (Phenomenex, Torrance, CA). Cacao fractions were prepared by sonicating (10 min) cacao tannin (10 mg) in a mixture of the mobile phase (1 mL) containing solvent A/B (75:25, v/v). Prior to HPLC injection the sample was centrifuged (16,100 g, 5 min) then filtered through a 0.45 lm PTFE filter. The binary mobile phase consisted of (A) acetonitrile/ acetic acid (98:2, v/v/v) and (B) methanol/water/acetic acid (95:3:2, v/v/v). HILIC separations were effected by a gradient of 0–40% solvent B from 0 to 60 min, 40–100% solvent B from 63 to 70 min, 100% isocratic B, and from 70 to 80 min, 100-0% B. Analysis was conducted with a 1.0 mL/min flow rate and a column temperature of 30 °C. Elution was monitored by DAD at 280 nm. For mass spectral analysis, 10 mM ammonium formate was added to the eluent post column via a tee (50 lL/min) to enhance ionisation efficiency. Ionisation was achieved by electrospray ionisation in the negative mode with a nebulizer at 50 psi, drying gas at 10 L/min and drying temperature at 350 °C. The ionisation parameters were segmented during the HPLC separation to optimise ionisation of cacao tannins with increasing molecular mass. The segmented ionisation parameters are described in Table 1. 2.5. Protein precipitation of cacao tannins 2.5.1. Increasing tannin size A 2 mg/mL stock solution of each cacao fraction was re-suspended in water. Aliquots from the stock solutions were then used to prepare tannin solutions in water of increasing concentration for each cacao fraction (50–2000 lg/mL). The iron reactive tannin content of each solution was determined against a standard curve prepared by reacting each corresponding tannin fraction with ferric chloride. The amount of iron reactive tannin precipitated in each solution by bovine serum albumin (BSA) protein was determined according to Harbertson, Kennedy, and Adams (2002) in triplicate. BSA has been used as model protein that mimics precipitation of salivary proteins with tannins that are connected to the sensation of astringency (Gawel, 1998; Green, 1993). Spectral analysis was conducted on a SpectraMax 384 UV–Vis absorbance microplate reader (Molecular Devices, Australia) using a polystyrene 96 well microplate (Greiner Bio-One, Interpath Services Pty. Ltd., West Heidelberg, Australia).
18
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Table 1 Segment parameters for ionisation of cacao tannins during HPLC analysis detected by electrospray ion trap mass spectrometry in negative ion mode. Segment (min)
DP
Target mass (m/z)
Scan range (m/z)
Compound stability (%)
Trap drive level (%)
4–16 16–24 24–38 38–42 42–46 46–70
1 2 3, 4 5 6 7+
289 577 1000 1441 1729 1008
200–600 200–1000 800–1200 1000–1500 1500–1750 900–2050
100 100 50 50 50 30
100 100 100 100 100 100
2.5.2. Tannin mixtures A known amount of each individual cacao fraction from tetramer to octamer was precipitated with BSA in combination with trimer at two different concentrations. The two concentrations of trimer were determined as being 1) 500 lg (a non-precipitable amount of trimer) and 2) 1000 lg (a precipitable amount of tannin). The total absorbance (510 nm) of iron reactive tannin precipitated with BSA was determined for each individual fraction and in combination with trimer at 500 lg and trimer at 1000 lg according to Harbertson et al. (2002) in duplicate.
concentration of BSA (1 mg/mL). The concentration of added monomer or tannin was plotted against the amount precipitated (Fig. 2). From trimer to octamer, the amount of tannin precipitated with protein (BSA) increased with increasing tannin concentration (Fig. 2). In addition, the amount of tannin precipitated increased with increasing size; 93% of the octamer added was precipitated (Table 2). As the size decreased the amount of precipitated tannin decreased with heptamers only precipitating around 64% of the added tannin, tetramers precipitating around 24% and trimers around 10% of the tannin added. Monomers and dimers did not precipitate with BSA at any concentration in the range tested here (50–2000 lg/mL).
2.6. Statistical analysis Means and standard errors for each data set were determined using Microsoft Excel and one-way analysis of variance (ANOVA) for tannin mixtures was conducted using Statistica 11.0 software (Statsoft Inc., USA).
3.3. Precipitation of mixed polymer systems The amount of tannin precipitated (measured by absorbance at 510 nm) with protein increased for each tannin fraction with the
3. Results 3.1. Qualitative analysis of cacao fractions Preparative diol phase HPLC separated cacao tannins into seven fractions of increasing molecular size (Fig. 1). Mass spectral data of collected fractions was in agreement with previously published studies (Hammerstone, Lazarus, Mitchell, Rucker, & Schmitz, 1999; Kelm et al., 2006). Condensed tannin fractions separated and eluted based on their degree of polymerisation from dimer to octamer. The purity of tannins from dimer to octamer was estimated to be 90–95%. Impurities were primarily lower or higher molecular weight tannins that preceded and followed the target tannin mass. 3.2. Precipitation of protein by tannins with increasing polymer size The monomer, epicatechin through to octamers was added at a range of concentrations (from 50 to 2000 lg/mL) to a fixed
Fig. 2. The concentration of each tannin added to BSA protein plotted against the concentration of tannin that was precipitated.
mAU
120
100
80
60
40
20
0 10
20
30
40
50
60
70
min
Fig. 1. Preparative diol phase HPLC chromatogram at 280 nm of cacao tannins separated into seven fractions of increasing molecular size.
19
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21 Table 2 Percentage of each tannin polymer precipitated by protein (BSA) and standard deviation (n = 3). The percentage of precipitated tannin was calculated from the determination of total tannin in the stock solution and re-dissolved precipitated tannin against a standard curve reacted with ferric chloride. Tannin polymer
% Tannin precipitated by BSA protein
Monomer Dimer Trimer Tetramer Pentamer Hexamer Heptamer Octamer
0 0 12.4 ± 0.02 23.9 ± 0.05 36.8 ± 0.02 32.7 ± 0.01 67.3 ± 0.01 93.4 ± 0.01
size up to octamer; around 93% of octamer was precipitated (Table 2). Following this trend, we would expect that at around 100% of decamer would be precipitated under these assay conditions. The logical conclusion from that observation would be that 100% of all tannins larger than decamer would be precipitated, although this needs to be tested; based on other systems it likely approaches asymptote beyond decamer. Despite a general trend of increasing protein precipitation with increasing tannin size, pentamer precipitated more than hexamer. The amount precipitated by both was similar and might be explained by differences in the conformation of pentamer and hexamer that affects the accessibility of binding sites on the hexamer. 4.2. Determination of a molar extinction coefficient for tannins
addition of trimers at both 500 lg and 1000 lg (Table 3). The increase in precipitation was significant for all tannin fractions with the addition of trimers at 1000 lg and for tetramers and hexamers with the addition of trimers at 500 lg.
4. Discussion 4.1. Relationship between condensed tannin size and protein precipitation The efficacy of tannin for precipitating protein is strongly influenced by tannin polymer size. The amount of tannin precipitated by protein increased with the addition of every subunit from trimers to octamers. It has previously been shown that a mixture of tannins with a high average size (degree of polymerisation) were more effective at precipitating protein than a mixture with a lower average size (Hagerman et al., 1998; Porter & Woodruffe, 1984). However, the presence of other structural features in the tannin used in those experiments made it unclear whether size or other structural characteristics were responsible for the increase in precipitation. The use of pure polymers in this study unequivocally demonstrates that polymer size increases the efficacy of protein precipitation within the range investigated here. Increased protein precipitation with increasing polymer size is primarily due to additional ortho di-hydroxy groups with every extra subunit. These provide more sites for hydrogen bonding to occur between the protein and tannin and increase opportunities for cross-linking between proteins. This increases the possibility of tannins and proteins forming larger aggregates that precipitate (McManus et al., 1985; Zanchi et al., 2008). The presence of galloylated subunits is also thought to increase hydrogen bonding and may have a stronger influence than size (McManus et al., 1985), but there has been no research to date that compares galloylation to polymer size. Galloylation might also increase hydrophobic interactions similar to pentagalloylglucose (Hagerman et al., 1998), but this also needs to be tested. This study demonstrated an increase in protein precipitation with increasing tannin size. Precipitation increased with increasing
A strong linear relationship was observed between the amount of tannin added and the amount precipitated, as well as between tannin size and the amount of tannin precipitated. This indicates that the method of quantification by reaction with ferric chloride and absorbance at 510 nm can accurately quantify individual tannin (epicatechin) subunits. Given this, the data collected here can be used to determine a calibration curve for tannin size based on absorbance of the ferric chloride reaction product, which in turn can be used to calculate the molar extinction coefficient of tannin-iron complexes from monomer to octamer (Fig. 3). While this will be useful in determining the concentration of pure polymers, it also demonstrates that the reaction between tannins and ferric chloride is proportional to the number of subunits in the polymer. It is also worth noting the relatively high absorptivity of monomer compared to the other polymers, which demonstrates the systematic error for quantification using this and other methods. While neither catechin nor epicatechin precipitate protein and are used to standardise analytical methods, the uniform reactivity of iron with catechin (and epicatechin) on a mass basis is suitable for the measurement of tannin of an unknown size and distribution by simply defining the mixture as the number of subunits present. 4.3. Impact of tannin mixtures on protein precipitation There is little to no data on mixed tannin systems of known composition in the literature. It was unclear whether the ability of tannin to precipitate protein would be altered when added as a mixture compared to individual tannin isolates. Given the differences in protein precipitation efficacy of different polymers reported here, it was considered possible that the ability of a small polymer, for example a trimer, to precipitate protein might be increased by the presence of a larger polymer, e.g. an octamer. In the present study we observed that mixed tannin systems precipitated protein primarily as an additive effect. When added to protein as a mixture containing both individual tannins (trimer to octamer) and trimers at a precipitable concentration (1000 lg), the iron reactive tannin measured in the mixed system was similar
Table 3 The amount of tannin polymer precipitated (PPT) by BSA protein individually and in mixtures with the addition of 500 and 1000 lg of trimer. Analysis was conducted in duplicate (n = 2) and shows standard deviation. Mean values within each row with the same letter(s) are not significantly different at p < 0.05.
Trimer (500 lg) Trimer (1000 lg) Tetramer (500 lg) Pentamer (500 lg) Hexamer (500 lg) Heptamer (300 lg) Octamer (300 lg)
Individually (lg CE)
+500 lg Trimer (lg CE)
+1000 lg Trimer (lg CE)
– 54.8 ± 1.15 51.7 ± 1.34c 109 ± 16.3b 178 ± 6.92c 118 ± 1.73b 146 ± 4.62b
– – 74.2 ± 0.77b 122 ± 2.89b 207 ± 16.2b 128 ± 1.73b 163 ± 14.6b
– – 111 ± 5.58a 173 ± 0.19a 251 ± 1.73a 207 ± 0.96a 199 ± 3.46a
20
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Fig. 3. Molar extinction coefficient of tannin polymers at 510 nm determined by their reaction with ferric chloride. Mean values (n = 3) for each polymer were significantly different at p < 0.05.
to the combined amount of tannin measured when the polymers were individually precipitated. This demonstrated that the individual tannins had the same efficacy for precipitating protein when combined as a mixture, an additive effect. However, when trimer was added to individual tannins (trimer to decamer) at a non-precipitable concentration (500 lg), we observed a small and in some cases significant increase in the amount of tannin precipitated. This increase in the amount of tannin precipitated as a mixture suggests that there might be a synergistic interaction occurring between tannins that enhance precipitation, although the increase in precipitation was generally insignificant compared to the overall additive effect of trimers added at the higher concentration. There are several possible explanations for the observed results. The increase in precipitation of tannin observed when trimer was added in combination with a larger tannin might be due to the availability of binding sites on the protein. It is possible that once the large tannin has saturated all accessible binding sites on the protein, there are still sites available for binding that the smaller tannin can access. Another possibility is that the larger tannin changes the shape of the protein structure during binding, which exposes more binding sites that trimers can form weak interactions with (e.g., hydrophobic or hydrophilic bonding). It is also possible that the synergistic effect occurs as a result of tannin aggregation at the higher concentration.
the effect of tannin size on astringency. The use of structurally pure tannins in this study provides the opportunity to assess size without the interference of other structural characteristics. In this study it was shown that monomers and dimers did not precipitate protein at all and that as little as 10% of trimer was precipitated, while up to 93% of octamer was precipitated. This indicates that a high proportion of lower molecular weight tannin does not precipitate protein and is therefore not contributing to perceived astringency; although low molecular weight tannins do contribute to bitterness (Casassa, Beaver, Mireles, & Harbertson, 2013a; Noble, 2002). As a result, methods that measure total tannin will over-estimate astringency because they effectively measure every subunit of every polymer regardless of size. Thus, in trying to measure astringency, what becomes important is the polymer length distribution (Hanlin, Kelm, Wilkinson, & Downey, 2011) and in particular the distribution of lower molecular weight tannins (Casassa et al., 2013b). Previous work has shown that astringency is correlated with tannin concentration (Kennedy et al., 2006; Landon, Weller, Harbertson, & Ross, 2008; Mercurio & Smith, 2008). If all tannin larger than octamer is fully precipitated, then above octamer, it is possible that only tannin concentration influences astringent perception. Simply, astringency is directly proportional to the number of tannin subunits (concentration) when those subunits comprise tannins larger than octamer. Having said that, we need to consider that in this work we have only looked at tannins from monomer to octamer with a single flavanol subunit (epicatechin) and we have used protein precipitation (defined as astringency) as a surrogate for perceived astringency, that is, mouthfeel, for example in wine. Our understanding of astringent perception and the correlation of protein precipitation with mouthfeel would benefit from sensory analysis of tannins over a range of polymer sizes such as those studied here. A descriptive sensory characterisation of non-precipitable tannins and those beyond octamers would also be helpful. Furthermore, it is unknown how astringent perception based on size is influenced by subunit modification such as esterification with gallic acid or addition of anthocyanin subunits or flavanols with different chemistry (e.g., catechin, gallocatechin, epigallocatechin). Astringent perception is thought to increase with galloylation and decrease with the addition of anthocyanins (Vidal et al., 2003, 2004), but it is unclear how this interacts with tannin size. Given the additional hydroxyl moieties, it is likely that galloylation will increase the efficacy of tannin for precipitating protein as it has been shown that seed tannin, which is high in galloyl substitution, was equal in astringency to skin tannin, which is low in galloylation and has larger polymers (Brossaud, Cheynier, & Noble, 2001).
4.4. Impact of condensed tannin size on protein precipitation and astringency 5. Conclusions Because tannins in most systems are a mixture of structures, they will not all have the same impact on protein precipitation and therefore on astringency. This has important implications in systems like red wine when trying to correlate measures of total tannin with expected astringent perception. Previous work has shown poor correlation between total tannin methods using gel permeation chromatography, phloroglucinolysis and protein precipitation and red wine astringency (Kennedy, Ferrier, Harbertson, & Peyrot des Gachons, 2006). A number of studies have investigated the impact of wine tannin content and structure on astringency by assessing different fractions and components of wine (Gonzalo-Diago, Dizy, & Fernández-Zurbano, 2013; Hufnagel & Hofmann, 2008; McRae, Schulkin, Kassara, Holt, & Smith, 2013; Wollmann & Hofmann, 2013). As these studies used structurally complex fractions that are not pure it is difficult to determine
Our hypotheses were that the efficacy of tannin for precipitating protein increases with size and the efficacy of polymer mixtures for precipitating protein is different to individual polymers. This study unequivocally demonstrated an increase in the efficacy of increasing size for precipitating protein within the range investigated here. This was achieved by the use of pure tannins without the interference of other structural characteristics such as galloylation. However, it is reasonable to postulate that an increase in size beyond octamers may not result in any further precipitation effectiveness. This study also demonstrated that mixtures of tannins primarily precipitate protein by an additive effect. A synergistic effect may occur in tannin mixtures to increase the amount of precipitable tannin at low tannin concentrations.
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Given the efficacy of tannin size and mixtures on protein precipitation, tannin size distribution is an important consideration for determining astringency and the use of total tannin measures for predicting astringency in wine. Acknowledgments This research was supported by the Victorian Department of Environment and Primary Industries, Washington Wine Grape Funds, Washington State University, Constellation Brands US and by the grape growers and winemakers of Australia through their investment body, the Grape and Wine Research and Development Corporation. References Adams, D. O., & Harbertson, J. F. (1999). Use of alkaline phosphatase for the analysis of tannins in grapes and red wines. American Journal of Enology and Viticulture, 50(3), 247–252. Brossaud, F., Cheynier, V., & Noble, A. C. (2001). Bitterness and astringency of grape and wine polyphenols. Australian Journal of Grape and Wine Research, 7, 33–39. Casassa, F. L., Beaver, C. W., Mireles, M. S., & Harbertson, J. F. (2013a). Effect of extended maceration and ethanol concentration on the extraction and evolution of phenolics, colour components and sensory attributes of Merlot wines. Australian Journal of Grape and Wine Research, 19, 25–39. Casassa, F. L., Larson, R. C., Beaver, C. W., Mireles, M. S., Keller, M., Riley, W. R., Smithyman, R., & Harbertson, J. F. (2013b). Impact of extended maceration and regulated deficit irrigation (RDI) in Cabernet Sauvignon wines: Characterization of proanthocyanidin distribution, anthocyanin extraction, and chromatic properties. Journal Agricultural and Food Chemistry, 61(26), 6446–6457. Charlton, A. J., Baxter, N. J., Lokman Khan, M., Moir, A. J. G., Haslam, E., Davies, A. P., & Williamson, M. P. (2002). Polyphenol/peptide binding and precipitation. Journal of Agricultural and Food Chemistry, 50, 1593–1601. Dixon, R. A., & Lamb, C. J. (1990). Molecular communications in interactions between plants and microbial pathogens. Annual Review of Plant Physiology and Plant Molecular Biology, 41, 339–367. Gawel, R. (1998). Red wine astringency: A review. Australian Journal of Grape and Wine Research, 4, 74–95. Gonzalo-Diago, A., Dizy, M., & Fernández-Zurbano, P. (2013). Taste and mouthfeel properties of red wines proanthocyanidins and their relation to the chemical composition. Journal of Agricultural and Food Chemistry, 61(37), 8861–8870. Green, B. G. (1993). Oral astringency: A tactile component of flavor. Acta Psychologica, 84(1), 119–125. Hagerman, A. E., & Butler, L. G. (1991). Tannins and Lignins. In G. A. Rosenthal & M. R. Berenbaum (Eds.), Herbivores: Their interactions with secondary plant metabolites, 2nd ed., vol. 1 (pp. 355–388). New York: Academic Press. Hagerman, A. E., Rice, M. E., & Ritchard, N. T. (1998). Mechanisms of protein precipitation for two tannins, pentagalloylglucose and epicatechin16(4–8) catechin (procyanidin). Journal of Agricultural and Food Chemistry, 46, 2590–2595. Hammerstone, J. F., Lazarus, A. S., Mitchell, A. E., Rucker, R., & Schmitz, H. H. (1999). Identification of procyanidins in cocoa (Theobroma cacao) and chocolate using high-performance liquid chromatography/mass spectrometry. Journal Agricultural and Food Chemistry, 47, 409–496. Hanlin, R. L., Kelm, M., Wilkinson, K. L., & Downey, M. O. (2011). Detailed characterization of proanthocyanidins in skin, seeds and wine of Shiraz and
21
Cabernet Sauvignon wine grapes (Vitis vinifera). Journal of Agricultural and Food Chemistry, 59(24), 13265–13276. Harbertson, J. F., Kennedy, J. A., & Adams, D. O. (2002). Tannin in skins and seeds of Cabernet Sauvignon, Syrah, and Pinot noir during ripening. American Journal of Enology and Viticulture, 53(1), 54–59. Harborne, J. B., & Grayer, R. J. (1993). Flavonoids and Insects. In J. B. Harborne (Ed.), The Flavonoids. Advances in Research since 1986, (pp. 589–618). London: Chapman & Hall. Haslam, E. (1998). Practical Polyphenolics: From Structure to Molecular Recognition and Physiological Action (1st ed.). Cambridge: Cambridge University Press. Hufnagel, J. C., & Hofmann, T. (2008). Orosensory-directed identification of astringent mouthfeel and bitter-tasting compounds in red wine. Journal of Agricultural and Food Chemistry, 56(4), 1376–1386. Kelm, M. A., Johnson, J. C., Robbins, R. J., Hammerstone, J. F., & Schmitz, H. H. (2006). High-Performance Liquid Chromatography separation and purification of cacao (Theobroma cacao L.) procyanidins according to degree of polymerization using a diol stationary phase. Journal of Agricultural and Food Chemistry, 54(5), 1571–1576. Kennedy, J. A., Ferrier, J., Harbertson, J. F., & Peyrot des Gachons, C. (2006). Analysis of tannins in red wine using multiple methods: Correlation with perceived astringency. American Journal of Enology and Viticulture, 57(4), 481–485. Landon, J. L., Weller, K., Harbertson, J. F., & Ross, C. F. (2008). Chemical and sensory evaluation of astringency in Washington State red wines. American Journal of Enology and Viticulture, 59(2), 153–158. Li, Y.-G., Tanner, G., & Larkin, P. (1996). The DMACA-HCl protocol and the threshold proanthocyanidin content for bloat safety in forage legumes. Journal of the Science of Food and Agriculture, 70(1), 89–101. Malien-Aubert, C., Dangles, O., & Amiot, M. J. (2002). Influence of procyanidins on the colour stability of oenin solutions. Journal of Agricultural and Food Chemistry, 50, 3299–3305. McManus, J. P., Davis, K. G., Beart, J. E., Gaffney, S. H., Lilley, T. H., & Haslam, E. (1985). Polyphenol interactions. Part 1. Introduction; some observations on the reversible complexation of polyphenols with proteins and polysaccharides. Journal of the Chemical Society, Perkin, Transactions 2. McRae, J. M., Schulkin, A., Kassara, S., Holt, H. E., & Smith, P. A. (2013). Sensory properties of wine tannin fractions: Implications for in-mouth sensory properties. Journal of Agricultural and Food Chemistry, 61(3), 719–727. Mercurio, M. D., & Smith, P. A. (2008). Tannin quantification in red grapes and wine: Comparison of polysaccharide- and protein-based tannin precipitation techniques and their ability to model wine astringency. Journal of Agricultural and Food Chemistry, 56(14), 5528–5537. Noble, A. C. (2002). Astringency and Bitterness of Flavonoid Phenols. In Chemistry of Taste, Vol. 825 (pp. 192–201): American Chemical Society. Porter, L. J., & Woodruffe, J. (1984). Haemanalysis: The realtive astringency of proanthocyanidin polymers. Phytochemistry, 23(6), 1255–1256. Ribéreau-Gayon, P. (1982). The anthocyanins of grapes and wines. In P. Markakis (Ed.), Anthocyanins as Food Colours (pp. 209–244). New York: Academic Press. Vidal, S., Francis, F., Guyot, S., Marnet, N., Kwiatkowski, M., Gawel, R., Cheynier, V., & Waters, E. J. (2003). The mouth-feel properties of grape and apple proanthocyanins in a wine-like medium. Journal of the Science of Food and Agriculture, 83, 564–573. Vidal, S., Francis, L., Noble, A., Kwiatkowski, M., Cheynier, V., & Waters, E. (2004). Taste and mouth-feel properties of different types of tannin-like polyphenolic compounds and anthocyanins in wine. Analytica Chimica Acta, 513, 57–65. Wollmann, N., & Hofmann, T. (2013). Compositional and sensory characterization of red wine polymers. Journal of Agricultural and Food Chemistry, 61(9), 2045–2061. Zanchi, D., Narayanan, T., Hagenmuller, D., Baron, A., Guyot, S., Cabane, B., & Bouhallab, S. (2008). Tannin-assisted aggregation of natively unfolded proteins. Europhysics Letters, 82(58001), 1–5.
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Impact of condensed tannin size as individual and mixed polymers on bovine serum albumin precipitation James F. Harbertson a, Rachel L. Kilmister b, Mark A. Kelm c, Mark O. Downey b,⇑ a
School of Food Science, Washington State University, Irrigated Agriculture Research and Extension Center, 24106 N. Bunn Rd., Prosser, WA 99350-8694, USA Department of Environment and Primary Industries, Victoria, PO Box 905, Mildura, Vic., Australia c Constellation Brands Inc., Mission Bell Winery, Research and Development, 12667 Road 24, Madera, CA 93637, USA b
a r t i c l e
i n f o
Article history: Received 18 October 2013 Received in revised form 4 March 2014 Accepted 8 March 2014 Available online 15 March 2014 Keywords: Condensed tannin Proanthocyanidin Theobroma cacao Protein precipitation Bovine serum albumin (BSA) Astringency
a b s t r a c t Condensed tannins composed of epicatechin from monomer to octamer were isolated from cacao (Theobroma cacao, L.) seeds and added to bovine serum albumin (BSA) individually and combined as mixtures. When added to excess BSA the amount of tannin precipitated increased with tannin size. The amount of tannin required to precipitate BSA varied among the polymers with the trimer requiring the most to precipitate BSA (1000 lg) and octamer the least (50 lg). The efficacy of condensed tannins for protein precipitation increased with increased degree of polymerisation (or size) from trimers to octamers (monomers and dimers did not precipitate BSA), while mixtures of two sizes primarily had an additive effect. This study demonstrates that astringent perception is likely to increase with increasing polymer size. Further research to expand our understanding of astringent perception and its correlation with protein precipitation would benefit from sensory analysis of condensed tannins across a range of polymer sizes. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Tannins are plant secondary metabolites that play a role in plant defence; mainly as a deterrent to herbivory (Hagerman & Butler, 1991; Harborne & Grayer, 1993), although there is some evidence of defence against microorganisms (Dixon & Lamb, 1990). Tannins also play an important role in a number of economically important areas of agricultural production; prevention of pasture bloat (Li, Tanner, & Larkin, 1996), the tanning of hides (Haslam, 1998) and in the mouthfeel of red wines (Gawel, 1998; Malien-Aubert, Dangles, & Amiot, 2002; Ribéreau-Gayon, 1982). In most of these systems, it is the ability of tannin to precipitate protein that underpins their function (Haslam, 1998). As a result the interaction of protein with tannin has attracted considerable research effort. Tannins are broadly grouped into two classes; hydrolysable and condensed tannins, both derived from plants. Hydrolysable tannins, which occur widely in wood, are based on gallic acid and occur as aggregations of gallates, known as ellagitannins, or are comprised of gallic acid moieties bound to a central glucose, e.g. penatgalloyl glucose and known as gallotannins (Haslam, 1998). Condensed tannins are polymers of flavanol subunits and are common in soft plant tissues such as leaves and fruit (skin and seeds)
⇑ Corresponding author. Tel.: +61 3 5051 4565; fax: +61 3 5051 4523. E-mail address: [email protected] (M.O. Downey). http://dx.doi.org/10.1016/j.foodchem.2014.03.026 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.
(Haslam, 1998). Tannins are a heterogeneous mixture containing a range of different polymer sizes, subunit compositions, and subunit linkages (Haslam, 1998). In grape seed alone it has been estimated that there is in excess of 65532 unique chemical tannin structures (Adams & Harbertson, 1999). Studies of tannin-protein interactions to date have given a strong indication that protein precipitation increases with increasing size (Porter & Woodruffe, 1984) and that the mode of interaction between tannins and proteins is a combination of hydrogen bonding and hydrophobic interactions (Charlton et al., 2002; Hagerman, Rice, & Ritchard, 1998). However, a limitation of all of the previous work was that owing to the complexity of isolating and purifying individual tannin polymers, these experiments have been conducted on mixtures of tannins with a range of polymer sizes and variable subunit composition (Porter & Woodruffe, 1984). The risk in using ill-defined mixtures and average polymer length to define tannin interactions and their correlation with astringency is the assumption that all tannins within a mixture interact with protein similarly. There have been no studies to date that document the interaction of thoroughly characterised individual condensed tannin polymers with protein. Consequently, several essential aspects about how tannins interact with protein are not well understood. There have been no studies of tannin-protein interactions that directly compare the protein precipitation of individual condensed tannins of increasing polymer size to unequivocally demonstrate the
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
relationship between polymer size and precipitation. The minimum molecular weight tannin required for protein precipitation has not yet been reported although it is thought that tannins less than tetramer do not precipitate bovine serum albumin protein (Adams & Harbertson, 1999). The impact on protein precipitation of structural features such as subunit composition is unclear as is the effect of different linkages (i.e. branched versus linear polymers). To address these questions, a methodology that would allow isolation of individual polymers was required. Such a method was developed for the separation of cacao polymers (Kelm, Johnson, Robbins, Hammerstone, & Schmitz, 2006). Condensed tannins from cacao are comprised of a single flavanol subunit, epicatechin, and can be readily separated up to octamer, although yields decrease with increasing polymer size. Using this method it has been possible to isolate sufficient quantity of pure polymers to determine the efficacy of tannin of different sizes for precipitating protein and the synergistic effect of these tannins in mixtures. Our hypotheses were that the efficacy of tannin for precipitating protein increases with size and the efficacy of polymer mixtures for precipitating protein is dependent upon polymer size.
17
liquid chromatography (HPLC) system using a Develosil diol column (300 50 mm i.d. 10 lm) in the hydrophilic interaction liquid chromatography (HILIC) mode. The binary mobile phase consisted of (A) acetonitrile and (B) methanol/water/formic acid (99:0.95:0.05, v/v/v). Gradient conditions for separation and collection were 0–30% solvent B from 0 to 35 min, 30% isocratic solvent B from 35 to 65 min, 30–80% solvent B from 65 to 67 min, and 80% isocratic solvent B from 67 to 80 min. Prior to injection the column was allowed to equilibrate (0% solvent B, 10 min). Cacao tannin (1 g) was dissolved in a 6 mL mixture of 25:75 mobile phase A/B, centrifuged (16,100g, 10 min), and filtered through a 0.45 lm PTFE filter prior to injection (5000 lL). The flow rate was set at 55 mL/min and the column temperature was maintained at room temperature (22 °C). Elution of separated cacao tannins was monitored by diode array detection (DAD) at 280 nm. Separated peaks were collected as they eluted, dried by rotary evaporation under reduced pressure (40 °C) then freeze dried to remove residual water and formic acid. Cacao fractions were prepared by this method three times for analysis. 2.4. HPLC-ESI-MS analysis of separated cacao tannins
2. Materials and methods 2.1. Chemicals Raw unfermented organic cacao (Theobroma cacao, L.) seeds of Equadorian origin were obtained from Natural Zing (Mt. Airy, MD, USA). Acetonitrile, methanol, hexane and ethyl acetate solvents were purchased from Merck (Darmstadt, Germany). Bovine serum albumin, sodium dodecyl sulphate, sodium hydroxide, triethanolamine, hydrochloric acid, acetic acid, formic acid, ferric chloride, ammonium formate, Amberlite resins, phosphorous pentoxide, ( )-epicatechin and (+)-catechin were purchased from Sigma-Aldrich Pty. Ltd (Castle Hill, Australia). HPLC columns were purchased from Phenomenex (Torrance, CA, USA). 2.2. Cacao tannin extraction and isolation Cacao seeds (500 g) were frozen with liquid nitrogen and ground in an electric coffee grinder to a 325 mesh (44 microns particle size). Ground cacao was defatted with hexane in a Soxhlet extractor for 6 hours. Defatted cacao was air dried overnight in a fume hood. To extract cacao tannin, a 100 g sample of defatted cacao was extracted twice with 1 L of 70% aqueous acetone on an orbital mixer (250 rpm; 30 min). To remove solids, the sample was filtered under vacuum through a glass microfiber filter. The filtrate was rotary evaporated under reduced pressure (40 °C) to remove acetone. The sample was mixed twice with 500 mL of ethyl acetate to remove low molecular weight material (epicatechin and tannin dimers). The aqueous phase was retained and mixed with Amberlite FPX62 resin (500 mg/mL) using an orbital mixer (250 rpm; 8 h). The sample was filtered under vacuum through a glass microfiber filter to remove the resin and the filtrate was then loaded on a glass column packed with Amberlite XAD7HP (20–60 mesh, 500 cm3) preconditioned with water (2 L). The column was washed with water (10 L) to remove any remaining anthocyanin, sugars and organic acids, then eluted with 80% aqueous ethanol (2 L). The eluate was rotary evaporated under reduced pressure (40 °C) to near dryness and further dried under vacuum with phosphorous pentoxide to yield a red-purple crusty solid (8.1 g). 2.3. Preparative HPLC fractionation of cacao tannin The fractionation of cacao tannin by degree of polymerisation was achieved on an Agilent 1100 preparative high performance
The qualitative analyses of separated cacao tannins was performed by HPLC (Agilent 1100) interfaced with an ion trap mass spectrometer (G2445D Bruker, Billerica, MA, USA). HPLC analysis was performed on a 250 4.6 mm i.d. 5 lm Develosil diol column fitted with a guard column of the same phase. (Phenomenex, Torrance, CA). Cacao fractions were prepared by sonicating (10 min) cacao tannin (10 mg) in a mixture of the mobile phase (1 mL) containing solvent A/B (75:25, v/v). Prior to HPLC injection the sample was centrifuged (16,100 g, 5 min) then filtered through a 0.45 lm PTFE filter. The binary mobile phase consisted of (A) acetonitrile/ acetic acid (98:2, v/v/v) and (B) methanol/water/acetic acid (95:3:2, v/v/v). HILIC separations were effected by a gradient of 0–40% solvent B from 0 to 60 min, 40–100% solvent B from 63 to 70 min, 100% isocratic B, and from 70 to 80 min, 100-0% B. Analysis was conducted with a 1.0 mL/min flow rate and a column temperature of 30 °C. Elution was monitored by DAD at 280 nm. For mass spectral analysis, 10 mM ammonium formate was added to the eluent post column via a tee (50 lL/min) to enhance ionisation efficiency. Ionisation was achieved by electrospray ionisation in the negative mode with a nebulizer at 50 psi, drying gas at 10 L/min and drying temperature at 350 °C. The ionisation parameters were segmented during the HPLC separation to optimise ionisation of cacao tannins with increasing molecular mass. The segmented ionisation parameters are described in Table 1. 2.5. Protein precipitation of cacao tannins 2.5.1. Increasing tannin size A 2 mg/mL stock solution of each cacao fraction was re-suspended in water. Aliquots from the stock solutions were then used to prepare tannin solutions in water of increasing concentration for each cacao fraction (50–2000 lg/mL). The iron reactive tannin content of each solution was determined against a standard curve prepared by reacting each corresponding tannin fraction with ferric chloride. The amount of iron reactive tannin precipitated in each solution by bovine serum albumin (BSA) protein was determined according to Harbertson, Kennedy, and Adams (2002) in triplicate. BSA has been used as model protein that mimics precipitation of salivary proteins with tannins that are connected to the sensation of astringency (Gawel, 1998; Green, 1993). Spectral analysis was conducted on a SpectraMax 384 UV–Vis absorbance microplate reader (Molecular Devices, Australia) using a polystyrene 96 well microplate (Greiner Bio-One, Interpath Services Pty. Ltd., West Heidelberg, Australia).
18
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Table 1 Segment parameters for ionisation of cacao tannins during HPLC analysis detected by electrospray ion trap mass spectrometry in negative ion mode. Segment (min)
DP
Target mass (m/z)
Scan range (m/z)
Compound stability (%)
Trap drive level (%)
4–16 16–24 24–38 38–42 42–46 46–70
1 2 3, 4 5 6 7+
289 577 1000 1441 1729 1008
200–600 200–1000 800–1200 1000–1500 1500–1750 900–2050
100 100 50 50 50 30
100 100 100 100 100 100
2.5.2. Tannin mixtures A known amount of each individual cacao fraction from tetramer to octamer was precipitated with BSA in combination with trimer at two different concentrations. The two concentrations of trimer were determined as being 1) 500 lg (a non-precipitable amount of trimer) and 2) 1000 lg (a precipitable amount of tannin). The total absorbance (510 nm) of iron reactive tannin precipitated with BSA was determined for each individual fraction and in combination with trimer at 500 lg and trimer at 1000 lg according to Harbertson et al. (2002) in duplicate.
concentration of BSA (1 mg/mL). The concentration of added monomer or tannin was plotted against the amount precipitated (Fig. 2). From trimer to octamer, the amount of tannin precipitated with protein (BSA) increased with increasing tannin concentration (Fig. 2). In addition, the amount of tannin precipitated increased with increasing size; 93% of the octamer added was precipitated (Table 2). As the size decreased the amount of precipitated tannin decreased with heptamers only precipitating around 64% of the added tannin, tetramers precipitating around 24% and trimers around 10% of the tannin added. Monomers and dimers did not precipitate with BSA at any concentration in the range tested here (50–2000 lg/mL).
2.6. Statistical analysis Means and standard errors for each data set were determined using Microsoft Excel and one-way analysis of variance (ANOVA) for tannin mixtures was conducted using Statistica 11.0 software (Statsoft Inc., USA).
3.3. Precipitation of mixed polymer systems The amount of tannin precipitated (measured by absorbance at 510 nm) with protein increased for each tannin fraction with the
3. Results 3.1. Qualitative analysis of cacao fractions Preparative diol phase HPLC separated cacao tannins into seven fractions of increasing molecular size (Fig. 1). Mass spectral data of collected fractions was in agreement with previously published studies (Hammerstone, Lazarus, Mitchell, Rucker, & Schmitz, 1999; Kelm et al., 2006). Condensed tannin fractions separated and eluted based on their degree of polymerisation from dimer to octamer. The purity of tannins from dimer to octamer was estimated to be 90–95%. Impurities were primarily lower or higher molecular weight tannins that preceded and followed the target tannin mass. 3.2. Precipitation of protein by tannins with increasing polymer size The monomer, epicatechin through to octamers was added at a range of concentrations (from 50 to 2000 lg/mL) to a fixed
Fig. 2. The concentration of each tannin added to BSA protein plotted against the concentration of tannin that was precipitated.
mAU
120
100
80
60
40
20
0 10
20
30
40
50
60
70
min
Fig. 1. Preparative diol phase HPLC chromatogram at 280 nm of cacao tannins separated into seven fractions of increasing molecular size.
19
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21 Table 2 Percentage of each tannin polymer precipitated by protein (BSA) and standard deviation (n = 3). The percentage of precipitated tannin was calculated from the determination of total tannin in the stock solution and re-dissolved precipitated tannin against a standard curve reacted with ferric chloride. Tannin polymer
% Tannin precipitated by BSA protein
Monomer Dimer Trimer Tetramer Pentamer Hexamer Heptamer Octamer
0 0 12.4 ± 0.02 23.9 ± 0.05 36.8 ± 0.02 32.7 ± 0.01 67.3 ± 0.01 93.4 ± 0.01
size up to octamer; around 93% of octamer was precipitated (Table 2). Following this trend, we would expect that at around 100% of decamer would be precipitated under these assay conditions. The logical conclusion from that observation would be that 100% of all tannins larger than decamer would be precipitated, although this needs to be tested; based on other systems it likely approaches asymptote beyond decamer. Despite a general trend of increasing protein precipitation with increasing tannin size, pentamer precipitated more than hexamer. The amount precipitated by both was similar and might be explained by differences in the conformation of pentamer and hexamer that affects the accessibility of binding sites on the hexamer. 4.2. Determination of a molar extinction coefficient for tannins
addition of trimers at both 500 lg and 1000 lg (Table 3). The increase in precipitation was significant for all tannin fractions with the addition of trimers at 1000 lg and for tetramers and hexamers with the addition of trimers at 500 lg.
4. Discussion 4.1. Relationship between condensed tannin size and protein precipitation The efficacy of tannin for precipitating protein is strongly influenced by tannin polymer size. The amount of tannin precipitated by protein increased with the addition of every subunit from trimers to octamers. It has previously been shown that a mixture of tannins with a high average size (degree of polymerisation) were more effective at precipitating protein than a mixture with a lower average size (Hagerman et al., 1998; Porter & Woodruffe, 1984). However, the presence of other structural features in the tannin used in those experiments made it unclear whether size or other structural characteristics were responsible for the increase in precipitation. The use of pure polymers in this study unequivocally demonstrates that polymer size increases the efficacy of protein precipitation within the range investigated here. Increased protein precipitation with increasing polymer size is primarily due to additional ortho di-hydroxy groups with every extra subunit. These provide more sites for hydrogen bonding to occur between the protein and tannin and increase opportunities for cross-linking between proteins. This increases the possibility of tannins and proteins forming larger aggregates that precipitate (McManus et al., 1985; Zanchi et al., 2008). The presence of galloylated subunits is also thought to increase hydrogen bonding and may have a stronger influence than size (McManus et al., 1985), but there has been no research to date that compares galloylation to polymer size. Galloylation might also increase hydrophobic interactions similar to pentagalloylglucose (Hagerman et al., 1998), but this also needs to be tested. This study demonstrated an increase in protein precipitation with increasing tannin size. Precipitation increased with increasing
A strong linear relationship was observed between the amount of tannin added and the amount precipitated, as well as between tannin size and the amount of tannin precipitated. This indicates that the method of quantification by reaction with ferric chloride and absorbance at 510 nm can accurately quantify individual tannin (epicatechin) subunits. Given this, the data collected here can be used to determine a calibration curve for tannin size based on absorbance of the ferric chloride reaction product, which in turn can be used to calculate the molar extinction coefficient of tannin-iron complexes from monomer to octamer (Fig. 3). While this will be useful in determining the concentration of pure polymers, it also demonstrates that the reaction between tannins and ferric chloride is proportional to the number of subunits in the polymer. It is also worth noting the relatively high absorptivity of monomer compared to the other polymers, which demonstrates the systematic error for quantification using this and other methods. While neither catechin nor epicatechin precipitate protein and are used to standardise analytical methods, the uniform reactivity of iron with catechin (and epicatechin) on a mass basis is suitable for the measurement of tannin of an unknown size and distribution by simply defining the mixture as the number of subunits present. 4.3. Impact of tannin mixtures on protein precipitation There is little to no data on mixed tannin systems of known composition in the literature. It was unclear whether the ability of tannin to precipitate protein would be altered when added as a mixture compared to individual tannin isolates. Given the differences in protein precipitation efficacy of different polymers reported here, it was considered possible that the ability of a small polymer, for example a trimer, to precipitate protein might be increased by the presence of a larger polymer, e.g. an octamer. In the present study we observed that mixed tannin systems precipitated protein primarily as an additive effect. When added to protein as a mixture containing both individual tannins (trimer to octamer) and trimers at a precipitable concentration (1000 lg), the iron reactive tannin measured in the mixed system was similar
Table 3 The amount of tannin polymer precipitated (PPT) by BSA protein individually and in mixtures with the addition of 500 and 1000 lg of trimer. Analysis was conducted in duplicate (n = 2) and shows standard deviation. Mean values within each row with the same letter(s) are not significantly different at p < 0.05.
Trimer (500 lg) Trimer (1000 lg) Tetramer (500 lg) Pentamer (500 lg) Hexamer (500 lg) Heptamer (300 lg) Octamer (300 lg)
Individually (lg CE)
+500 lg Trimer (lg CE)
+1000 lg Trimer (lg CE)
– 54.8 ± 1.15 51.7 ± 1.34c 109 ± 16.3b 178 ± 6.92c 118 ± 1.73b 146 ± 4.62b
– – 74.2 ± 0.77b 122 ± 2.89b 207 ± 16.2b 128 ± 1.73b 163 ± 14.6b
– – 111 ± 5.58a 173 ± 0.19a 251 ± 1.73a 207 ± 0.96a 199 ± 3.46a
20
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Fig. 3. Molar extinction coefficient of tannin polymers at 510 nm determined by their reaction with ferric chloride. Mean values (n = 3) for each polymer were significantly different at p < 0.05.
to the combined amount of tannin measured when the polymers were individually precipitated. This demonstrated that the individual tannins had the same efficacy for precipitating protein when combined as a mixture, an additive effect. However, when trimer was added to individual tannins (trimer to decamer) at a non-precipitable concentration (500 lg), we observed a small and in some cases significant increase in the amount of tannin precipitated. This increase in the amount of tannin precipitated as a mixture suggests that there might be a synergistic interaction occurring between tannins that enhance precipitation, although the increase in precipitation was generally insignificant compared to the overall additive effect of trimers added at the higher concentration. There are several possible explanations for the observed results. The increase in precipitation of tannin observed when trimer was added in combination with a larger tannin might be due to the availability of binding sites on the protein. It is possible that once the large tannin has saturated all accessible binding sites on the protein, there are still sites available for binding that the smaller tannin can access. Another possibility is that the larger tannin changes the shape of the protein structure during binding, which exposes more binding sites that trimers can form weak interactions with (e.g., hydrophobic or hydrophilic bonding). It is also possible that the synergistic effect occurs as a result of tannin aggregation at the higher concentration.
the effect of tannin size on astringency. The use of structurally pure tannins in this study provides the opportunity to assess size without the interference of other structural characteristics. In this study it was shown that monomers and dimers did not precipitate protein at all and that as little as 10% of trimer was precipitated, while up to 93% of octamer was precipitated. This indicates that a high proportion of lower molecular weight tannin does not precipitate protein and is therefore not contributing to perceived astringency; although low molecular weight tannins do contribute to bitterness (Casassa, Beaver, Mireles, & Harbertson, 2013a; Noble, 2002). As a result, methods that measure total tannin will over-estimate astringency because they effectively measure every subunit of every polymer regardless of size. Thus, in trying to measure astringency, what becomes important is the polymer length distribution (Hanlin, Kelm, Wilkinson, & Downey, 2011) and in particular the distribution of lower molecular weight tannins (Casassa et al., 2013b). Previous work has shown that astringency is correlated with tannin concentration (Kennedy et al., 2006; Landon, Weller, Harbertson, & Ross, 2008; Mercurio & Smith, 2008). If all tannin larger than octamer is fully precipitated, then above octamer, it is possible that only tannin concentration influences astringent perception. Simply, astringency is directly proportional to the number of tannin subunits (concentration) when those subunits comprise tannins larger than octamer. Having said that, we need to consider that in this work we have only looked at tannins from monomer to octamer with a single flavanol subunit (epicatechin) and we have used protein precipitation (defined as astringency) as a surrogate for perceived astringency, that is, mouthfeel, for example in wine. Our understanding of astringent perception and the correlation of protein precipitation with mouthfeel would benefit from sensory analysis of tannins over a range of polymer sizes such as those studied here. A descriptive sensory characterisation of non-precipitable tannins and those beyond octamers would also be helpful. Furthermore, it is unknown how astringent perception based on size is influenced by subunit modification such as esterification with gallic acid or addition of anthocyanin subunits or flavanols with different chemistry (e.g., catechin, gallocatechin, epigallocatechin). Astringent perception is thought to increase with galloylation and decrease with the addition of anthocyanins (Vidal et al., 2003, 2004), but it is unclear how this interacts with tannin size. Given the additional hydroxyl moieties, it is likely that galloylation will increase the efficacy of tannin for precipitating protein as it has been shown that seed tannin, which is high in galloyl substitution, was equal in astringency to skin tannin, which is low in galloylation and has larger polymers (Brossaud, Cheynier, & Noble, 2001).
4.4. Impact of condensed tannin size on protein precipitation and astringency 5. Conclusions Because tannins in most systems are a mixture of structures, they will not all have the same impact on protein precipitation and therefore on astringency. This has important implications in systems like red wine when trying to correlate measures of total tannin with expected astringent perception. Previous work has shown poor correlation between total tannin methods using gel permeation chromatography, phloroglucinolysis and protein precipitation and red wine astringency (Kennedy, Ferrier, Harbertson, & Peyrot des Gachons, 2006). A number of studies have investigated the impact of wine tannin content and structure on astringency by assessing different fractions and components of wine (Gonzalo-Diago, Dizy, & Fernández-Zurbano, 2013; Hufnagel & Hofmann, 2008; McRae, Schulkin, Kassara, Holt, & Smith, 2013; Wollmann & Hofmann, 2013). As these studies used structurally complex fractions that are not pure it is difficult to determine
Our hypotheses were that the efficacy of tannin for precipitating protein increases with size and the efficacy of polymer mixtures for precipitating protein is different to individual polymers. This study unequivocally demonstrated an increase in the efficacy of increasing size for precipitating protein within the range investigated here. This was achieved by the use of pure tannins without the interference of other structural characteristics such as galloylation. However, it is reasonable to postulate that an increase in size beyond octamers may not result in any further precipitation effectiveness. This study also demonstrated that mixtures of tannins primarily precipitate protein by an additive effect. A synergistic effect may occur in tannin mixtures to increase the amount of precipitable tannin at low tannin concentrations.
J.F. Harbertson et al. / Food Chemistry 160 (2014) 16–21
Given the efficacy of tannin size and mixtures on protein precipitation, tannin size distribution is an important consideration for determining astringency and the use of total tannin measures for predicting astringency in wine. Acknowledgments This research was supported by the Victorian Department of Environment and Primary Industries, Washington Wine Grape Funds, Washington State University, Constellation Brands US and by the grape growers and winemakers of Australia through their investment body, the Grape and Wine Research and Development Corporation. References Adams, D. O., & Harbertson, J. F. (1999). Use of alkaline phosphatase for the analysis of tannins in grapes and red wines. American Journal of Enology and Viticulture, 50(3), 247–252. Brossaud, F., Cheynier, V., & Noble, A. C. (2001). Bitterness and astringency of grape and wine polyphenols. Australian Journal of Grape and Wine Research, 7, 33–39. Casassa, F. L., Beaver, C. W., Mireles, M. S., & Harbertson, J. F. (2013a). Effect of extended maceration and ethanol concentration on the extraction and evolution of phenolics, colour components and sensory attributes of Merlot wines. Australian Journal of Grape and Wine Research, 19, 25–39. Casassa, F. L., Larson, R. C., Beaver, C. W., Mireles, M. S., Keller, M., Riley, W. R., Smithyman, R., & Harbertson, J. F. (2013b). Impact of extended maceration and regulated deficit irrigation (RDI) in Cabernet Sauvignon wines: Characterization of proanthocyanidin distribution, anthocyanin extraction, and chromatic properties. Journal Agricultural and Food Chemistry, 61(26), 6446–6457. Charlton, A. J., Baxter, N. J., Lokman Khan, M., Moir, A. J. G., Haslam, E., Davies, A. P., & Williamson, M. P. (2002). Polyphenol/peptide binding and precipitation. Journal of Agricultural and Food Chemistry, 50, 1593–1601. Dixon, R. A., & Lamb, C. J. (1990). Molecular communications in interactions between plants and microbial pathogens. Annual Review of Plant Physiology and Plant Molecular Biology, 41, 339–367. Gawel, R. (1998). Red wine astringency: A review. Australian Journal of Grape and Wine Research, 4, 74–95. Gonzalo-Diago, A., Dizy, M., & Fernández-Zurbano, P. (2013). Taste and mouthfeel properties of red wines proanthocyanidins and their relation to the chemical composition. Journal of Agricultural and Food Chemistry, 61(37), 8861–8870. Green, B. G. (1993). Oral astringency: A tactile component of flavor. Acta Psychologica, 84(1), 119–125. Hagerman, A. E., & Butler, L. G. (1991). Tannins and Lignins. In G. A. Rosenthal & M. R. Berenbaum (Eds.), Herbivores: Their interactions with secondary plant metabolites, 2nd ed., vol. 1 (pp. 355–388). New York: Academic Press. Hagerman, A. E., Rice, M. E., & Ritchard, N. T. (1998). Mechanisms of protein precipitation for two tannins, pentagalloylglucose and epicatechin16(4–8) catechin (procyanidin). Journal of Agricultural and Food Chemistry, 46, 2590–2595. Hammerstone, J. F., Lazarus, A. S., Mitchell, A. E., Rucker, R., & Schmitz, H. H. (1999). Identification of procyanidins in cocoa (Theobroma cacao) and chocolate using high-performance liquid chromatography/mass spectrometry. Journal Agricultural and Food Chemistry, 47, 409–496. Hanlin, R. L., Kelm, M., Wilkinson, K. L., & Downey, M. O. (2011). Detailed characterization of proanthocyanidins in skin, seeds and wine of Shiraz and
21
Cabernet Sauvignon wine grapes (Vitis vinifera). Journal of Agricultural and Food Chemistry, 59(24), 13265–13276. Harbertson, J. F., Kennedy, J. A., & Adams, D. O. (2002). Tannin in skins and seeds of Cabernet Sauvignon, Syrah, and Pinot noir during ripening. American Journal of Enology and Viticulture, 53(1), 54–59. Harborne, J. B., & Grayer, R. J. (1993). Flavonoids and Insects. In J. B. Harborne (Ed.), The Flavonoids. Advances in Research since 1986, (pp. 589–618). London: Chapman & Hall. Haslam, E. (1998). Practical Polyphenolics: From Structure to Molecular Recognition and Physiological Action (1st ed.). Cambridge: Cambridge University Press. Hufnagel, J. C., & Hofmann, T. (2008). Orosensory-directed identification of astringent mouthfeel and bitter-tasting compounds in red wine. Journal of Agricultural and Food Chemistry, 56(4), 1376–1386. Kelm, M. A., Johnson, J. C., Robbins, R. J., Hammerstone, J. F., & Schmitz, H. H. (2006). High-Performance Liquid Chromatography separation and purification of cacao (Theobroma cacao L.) procyanidins according to degree of polymerization using a diol stationary phase. Journal of Agricultural and Food Chemistry, 54(5), 1571–1576. Kennedy, J. A., Ferrier, J., Harbertson, J. F., & Peyrot des Gachons, C. (2006). Analysis of tannins in red wine using multiple methods: Correlation with perceived astringency. American Journal of Enology and Viticulture, 57(4), 481–485. Landon, J. L., Weller, K., Harbertson, J. F., & Ross, C. F. (2008). Chemical and sensory evaluation of astringency in Washington State red wines. American Journal of Enology and Viticulture, 59(2), 153–158. Li, Y.-G., Tanner, G., & Larkin, P. (1996). The DMACA-HCl protocol and the threshold proanthocyanidin content for bloat safety in forage legumes. Journal of the Science of Food and Agriculture, 70(1), 89–101. Malien-Aubert, C., Dangles, O., & Amiot, M. J. (2002). Influence of procyanidins on the colour stability of oenin solutions. Journal of Agricultural and Food Chemistry, 50, 3299–3305. McManus, J. P., Davis, K. G., Beart, J. E., Gaffney, S. H., Lilley, T. H., & Haslam, E. (1985). Polyphenol interactions. Part 1. Introduction; some observations on the reversible complexation of polyphenols with proteins and polysaccharides. Journal of the Chemical Society, Perkin, Transactions 2. McRae, J. M., Schulkin, A., Kassara, S., Holt, H. E., & Smith, P. A. (2013). Sensory properties of wine tannin fractions: Implications for in-mouth sensory properties. Journal of Agricultural and Food Chemistry, 61(3), 719–727. Mercurio, M. D., & Smith, P. A. (2008). Tannin quantification in red grapes and wine: Comparison of polysaccharide- and protein-based tannin precipitation techniques and their ability to model wine astringency. Journal of Agricultural and Food Chemistry, 56(14), 5528–5537. Noble, A. C. (2002). Astringency and Bitterness of Flavonoid Phenols. In Chemistry of Taste, Vol. 825 (pp. 192–201): American Chemical Society. Porter, L. J., & Woodruffe, J. (1984). Haemanalysis: The realtive astringency of proanthocyanidin polymers. Phytochemistry, 23(6), 1255–1256. Ribéreau-Gayon, P. (1982). The anthocyanins of grapes and wines. In P. Markakis (Ed.), Anthocyanins as Food Colours (pp. 209–244). New York: Academic Press. Vidal, S., Francis, F., Guyot, S., Marnet, N., Kwiatkowski, M., Gawel, R., Cheynier, V., & Waters, E. J. (2003). The mouth-feel properties of grape and apple proanthocyanins in a wine-like medium. Journal of the Science of Food and Agriculture, 83, 564–573. Vidal, S., Francis, L., Noble, A., Kwiatkowski, M., Cheynier, V., & Waters, E. (2004). Taste and mouth-feel properties of different types of tannin-like polyphenolic compounds and anthocyanins in wine. Analytica Chimica Acta, 513, 57–65. Wollmann, N., & Hofmann, T. (2013). Compositional and sensory characterization of red wine polymers. Journal of Agricultural and Food Chemistry, 61(9), 2045–2061. Zanchi, D., Narayanan, T., Hagenmuller, D., Baron, A., Guyot, S., Cabane, B., & Bouhallab, S. (2008). Tannin-assisted aggregation of natively unfolded proteins. Europhysics Letters, 82(58001), 1–5.
