Abstract
A 96-well microplate reader was found to be an efficient and cost-effective tool for assaying red wines using an existing comprehensive red wine phenolics assay. This instrument increases the throughput three-fold while also reducing the time required to complete the assay. A traditional UV-vis capable spectrophotometer was used as the reference method to validate the procedure for the analysis of tannin, iron-reactive phenolics, anthocyanin, large polymeric pigment, and small polymeric pigment. The iron-reactive phenolics and anthocyanin steps were completely scaled down to microplate volume. The tannin and polymeric pigment measurements required two initial steps in microfuge tubes before transfer to the microplate. The 40 wine samples representing four different wine types were assayed and compared using the two different instruments. The average instrumental differences between microplate and spectrophotometer ranged from 1 to 6% for tannin, 1 to 5% for iron-reactive phenolics, <1 to 4% for anthocyanin, and 0.50 to 6.66% for large and small polymeric pigments. The largest discrepancies were seen in the large polymeric pigment measurements, where the greatest difference was 6.66% for the 2003 Merlot samples.
In high-throughput winery laboratories it would be advantageous to have a rapid, efficient, and cost-effective assay for the precise measurement of phenolic components in wine. A μQuant 96-well microplate reader was evaluated for use with a comprehensive red wine phenolics assay (Harbertson et al. 2003), which was previously optimized using a traditional UV-vis capable spectrophotometer. As with the traditional spectrophotometer, the assay on a microplate reader uses vis-capable spectrophotometry, protein precipitation, and bisulfite bleaching techniques (Picciotto 2002) to measure tannin, iron-reactive phenolics, anthocyanin, and polymeric pigment in red wines. In this version of the assay, the iron-reactive phenolics and anthocyanin steps are completely scaled down to microplate volume. The tannin and polymeric pigment measurements require that two initial steps are set up in microfuge tubes before transfer to the microplate.
This assay was incorporated into the laboratory procedures at Joseph Phelps Vineyards on an experimental level in 2002 and was run on a basic, single-cell spectrophotometer. It quickly became apparent that the data obtained was critical to winemaking; especially with respect to fermentation, postfermentation maceration, pressing, and blending decisions. The assay was found to be time-consuming (2 to 3 hr for 12 samples) and several modifications were made with the goal of streamlining the procedure to reduce turnaround time during harvest. The most recent development was the purchase of a microplate reader in spring 2004. This instrument has allowed a three-fold or greater increase in the throughput of the assay (2 to 3 hr for 36 samples) and the new procedure has reduced reagent costs five-fold. Instead of preparing and reading one sample at a time, the user can prepare eight samples simultaneously using a multichannel pipettor and can read up to 48 samples in duplicate on the plate reader. Labor and reagent costs are significantly reduced when using this improved method, and waste is greatly reduced by eliminating the use of disposable cuvettes and minimizing the quantity of 1.5-mL microfuge tubes needed for the assay.
A rapid and efficient phenolics assay in the winery laboratory allows winemakers and enologists to make quick and reliable decisions on fermentation, pressing, and blending on the same day a sample is submitted for analysis. The efficiency of the assay on the microplate reader is such that, in an average-size winery (50,000 to 80,000 cases annually), all fermentations could be monitored for phenolic development during the critical maceration periods and each pressing decision could be made with the confidence that the desired extraction has been achieved. Although a detailed discussion of fermentation data is beyond the scope of this article, it is worthwhile to note that some of the phenolic components were of particular interest: tannin, iron-reactive phenolics, and anthocyanin values were found to be the best indicators of the level of extraction in any given fermentation. Elimination of the large polymeric pigment and small polymeric pigment steps of the assay yielded a significant time savings, allowing a greater number of samples to be assayed.
Materials and Methods
Wine samples.
The evaluation included several wine types and represented a wide range of phenolic concentrations, with Pinot noir being the lowest and Cabernet Sauvignon blends the highest. The Syrah blends represented high anthocyanin levels and Merlot was included in the study because it is a good industry standard. Sample dilution using a model wine solution, buffer B (Table 1⇓), ranged from 1 to 5 for Cabernet Sauvignon, Merlot, and Syrah to no dilution for Pinot noir. The commercial wines included in the study were Cabernet Sauvignon blends and Merlot from Napa Valley, Pinot noir from California (one from Burgundy, France), and Syrah from Monterey and Napa Valley.
Preparation of reagents for comprehensive red wine phenolics assay (summarized from Harbertson et al. 2003, Picciotto 2002).
Microplate reader path-length calibration.
In order to model the original procedure, which was run on a traditional spectrophotometer with a 1-cm path length, it was necessary to calibrate the path length on the μQuant microplate reader (Bio-Tek Instruments, Winooski, VT). The calculation of concentration from absorbance using extinction coefficients requires that the path length of the absorbing material be known. Horizontal photometry, as performed in spectrophotometers, uses physical dimensions of the cuvette to fix the light path length of the sample. In vertical photometry, as performed by microplate readers, the light passes through the vertical axis of the absorbing solution. As a result, the path length of the absorbing solution is dependent on the volume of the solution in the well. In order to use absorbance data from vertical photometric devices to calculate concentrations, absorbance measurements need to be normalized to 1 cm. The path length of a solution in a microplate well can be determined by comparing the absorbance of a dye solution in a 1-cm cuvette to the absorbance in the microplate well. The ratio of the microplate well absorbance value to the 1-cm value is equal to the path length in centimeters. Water has a small yet significant peak in absorbance at 977 nm (McGowan et al. 1997): the ratio of the absorbance of a water sample at 977 nm in a microplate well to that of water in a 1-cm cuvette is equivalent to the path length.
The path length of each different reaction volume used in the assay was determined experimentally before performing the analyses described below, and a reference wavelength of 900 nm was used to subtract any background absorbance caused by the microplate itself. Once the path length for each reaction volume had been determined, the values were used in the KC4 software (Bio-Tek Instruments) protocol to compute concentrations from absorbance values. The software was set to automatically divide the raw absorbance values by the corresponding determined path length, which was entered as a constant, to normalize the data to 1 cm. A separate protocol in the KC4 software was established for each routine absorbance measurement in the assay so that these calculations could automatically be performed for each measurement.
Reagents.
Maleic acid and potassium metabisulfite were purchased from Fisher Scientific (Fair Lawn, NJ). Bovine serum albumin (98%; Cohn Fraction V; molecular weight ~66 kDa), sodium dodecyl sulfate (SDS; lauryl sulfate, sodium salt), triethanolamine, ferric chloride hexahydrate (97%), (+)-catechin, and all other buffer reagents were purchased from Sigma (St. Louis, MO). Reagents were prepared as indicated in (Harbertson et al. 2003). Reagent preparation and storage conditions are summarized in Table 1⇑.
Catechin standard curve.
Catechin solution was prepared in 50-mL total volume and then aliquotted into 5-mL borosilicate tubes (Fisher Scientific) that were capped, sealed, and then frozen for no longer than 60 days (catechin degradation occurred with longer storage times). Standard curve preparation was scaled down from that described elsewhere (Harbertson et al. 2002) in order to adapt to the microplate reader. The standard curve was prepared in 14 wells of a clean polystyrene microplate (Corning, Inc., Corning, NY). A new set of tips was used between each volume transfer into the microplate, as it significantly improved reproducibility within duplicate sets on the microplate reader. The practice of changing to new tips between each volume transfer into the microplate was followed throughout the microplate assay. Two wells each received the following volumes of catechin standard solution: 0, 15, 30, 45, 60, 75, and 90 μL (concentration range = 0 to 300 mg/L catechin). The volume in each well was adjusted to 262 μL with buffer C and mixed by vortexing the plate at “shake, level 2” using a foam insert (Multiple Sample Head Starter Set; Fisher Scientific). A volume of 38μL of ferric chloride solution was added to each well, then mixed and incubated 10 min, followed by absorbance determination at 510 nm, where buffer C was used as the blank. A graph was plotted in Excel (version 11.5612; Microsoft, Redmond, WA) and linear regression analysis was used to determine the amount of tannin and iron-reactive phenolics in each sample.
Polymeric pigment analysis.
The method of Harbertson et al. (2003) was scaled down and volumes were adjusted to adapt to the microplate reader. Wine samples were diluted (1:5 for Cabernet Sauvignon, Merlot, and Syrah; no dilution for Pinot noir) in buffer B and each sample was assayed in duplicate. Simultaneous analysis of polymeric pigment and tannin requires two 1.5-mL microfuge tubes and a 96-well microplate. The first step was carried out in a 1.5-mL microfuge tube and transferred to the microplate, as reaction dynamics for measurement A were more consistent when set up in full-scale volume than in microplate volume. A volume of 500 μL of diluted wine sample was added to a 1.5-mL microfuge tube followed by the addition of 1 mL of buffer A and immediate mixing by inversion. Next, a 300-μL aliquot of this mixture was transferred to a microplate and incubated at room temperature for 10 min. The absorbance at 520 nm was determined for each well (measurement A). Using a Transferpette-8, 5- to 50-μL multichannel pipettor (BrandTech Scientific, Essex, CT), 24 μL of bleaching solution was added to each well and then mixed by pipetting followed by 10-min incubation. Absorbance was determined at 520 nm (measurement B).
The next polymeric pigment step required set up of the tannin analysis portion of the assay: tannin analysis was started in microfuge tubes and then transferred to a microplate. Because the tannin analysis requires centrifugation at 13,500 g to pellet the tannin-protein precipitate, this step was performed in 1.5-mL microfuge tubes according to (Harbertson et al. 2002). After centrifugation, a 300-μL aliquot of supernatant from each reaction tube was transferred to a microplate in duplicate. A volume of 24 μL of bleaching solution was then added to each well, followed by mixing, 10-min incubation, and then absorbance was determined at 520 nm (measurement C). For each of the above measurements, the microplate reader was blanked using buffer A. The remaining supernatant was carefully removed using a vacuum apparatus, with a 200-μL pipette tip attached, and then the steps for pellet treatment were carried out (Harbertson et al. 2002). Measurement C represents small polymeric pigment (SPP), and measurement B minus measurement C represents large polymeric pigment (LPP) (Harbertson et al. 2002). Measurement C (SPP) was multiplied by a factor of 10/7 to account for a fraction of SPP that bleaches with bisulfite, while the LPP difference measurement was multiplied by a factor of 4/3 to account for LPP bleaching. These factors were established empirically (Picciotto 2002). A factor of 1.08 was used to account for the addition of bleaching solution and a factor of 3 to account for reaction dilution. LPP and SPP were calculated as follows: LPP = B-C and SPP = C. For example, LPP = (1.08)*(dilution factor)*(4/3)*(B-C) and SPP = (1.08)*(dilution factor)*(10/7)*(C). Assay calculations are summarized in Table 2⇓.
Summary of calculations for comprehensive red wine phenolics assay (from Harbertson et al. 2003, Picciotto 2002).
Tannin analysis.
After the protein-tannin pellet was redissolved in buffer C (Harbertson et al. 2002), a 262-μL aliquot of this mixture was transferred to a microplate, and the absorbance at 510 nm was determined immediately (tannin background measurement). Then 38 μL of ferric chloride reagent was added to each well and mixed by pipetting using an electronic Transferpette-8/12, 30 to 300-μL multichannel pipettor (BrandTech Scientific). This reaction was incubated 10 min, followed by absorbance determination at 510 nm (tannin final measurement). For each of the above measurements, the microplate reader was blanked with buffer C. The absorbance due to tannin in the wine sample was calculated as described elsewhere (Harbertson et al. 2003), and then the amount of tannin in each sample was determined from a catechin standard curve as described above. The amount of tannin in the sample, expressed in mg/L catechin equivalents, was calculated from the tannin final absorbance minus the tannin background absorbance. “Absorbance due tannin” = [(Tannin Final @510 nm) − (0 mg/L catechin @510nm)]-(Tannin Background @ 510 nm * 0.875). Using the catechin standard curve: tannin = 2 * [(Abs due tannin − intercept)/slope]*(dilution factor), where the factor “2” accounts for reaction dilution.
Iron-reactive phenolics analysis.
The iron-reactive phenolics (IRP) portion of the assay was scaled down from the 1.0-mL reaction volume described elsewhere (Harbertson et al. 2004). A volume of 15 μL of undiluted wine sample (to give final dilution of 1:20) was added per well and then 247 μL of buffer C was added, followed by pipette mixing, 10-min incubation, and absorbance determination at 510 nm (IRP background measurement). To each well, 38 μL of ferric chloride reagent was added, followed by mixing, 10-min incubation, and absorbance determination at 510 nm (IRP final measurement). The microplate reader was blanked with buffer C for each of the above measurements. The amount of IRP in each wine sample was determined as described above for tannin using the difference of IRP final and background measurements, followed by linear regression analysis from the catechin standard curve and a dilution factor of 20 for all samples. Iron-reactive phenolics in wine samples are expressed in mg/L catechin equivalents.
Anthocyanin analysis.
A simple method for anthocyanin analysis in wine is described elsewhere (Picciotto 2002). In our work, the volumes were scaled down to adapt to the microplate reader. A volume of 50 μL of each diluted wine sample was added per well, followed by 50 μL of buffer B to double the dilution, and then each well was mixed by pipetting. This extra dilution step was necessary because of high-color samples that frequently absorbed above 2.0 absorbance units at a 1:5 dilution. A volume of 200 μL of buffer D was then added to each well, followed by mixing and 10-min incubation, and then absorbance was determined at 520 nm (measurement D). The amount of anthocyanin in the wine sample is represented in malvidin-3-glucoside (M-3-G) units, where the absorbance units are converted into M-3-G units using a coefficient of 0.0102, as determined by Picciotto (2002) under the conditions of this assay. The amount of anthocyanin in the sample, expressed in mg/L M-3-G, was calculated from the measurement D absorbance minus measurement A absorbance and then dividing that difference by 0.0102. Using Syrah as an example, the dilution factor for measurement A was 5 and the dilution factor for measurement D was 10 (because the dilution in this step was doubled): anthocyanin (mg/L M-3-G) = [(Measurement D * 10) − (Measurement A * 5)]/0.0102.
Results and Discussion
Forty wine samples were assayed, representing a range of phenolic compositions including Pinot noir, Merlot, Syrah, and Cabernet Sauvigon. The scaled-down microplate procedure described above was validated by comparing assay values for each wine to the full-scale method described elsewhere (Harbertson et al. 2002, 2003, 2004, Picciotto 2002). The relationship between assay values obtained using the two instruments and the tannin, IRP, and anthocyanin for each wine type are shown in Figure 1⇓.
Comparison of MicroQuant plate reader (μQuant) and spectrophotometer (Spec) assay averages for commercial red wines as determined by the comprehensive red wine phenolics assay (Harbertson et al. 2003).
Previous research has shown that tannin and other polyphenols adsorb onto polystyrene (Lam et al. 1970, Gray 1978, Oh et. al 1980, Matsuo et. al 1995). However, there is evidence that ionic bonding of proanthocyanidins to a polystyrene surface is excluded by the absence of charged groups on proanthocyanidins at pH values considerably below the pKa of the phenolic groups (Oh et al. 1980). The low pH in the anthocyanin and polymeric pigment steps of the assay are consistent with the prevention of significant binding to the polystyrene plate. Our small average differences between methods as discussed below show that we did not observe any significant interference by polyphenol adsorption to the polystyrene plate.
Validation was established by calculating instrumental differences for each wine type after running the full assay on the microplate reader and then a spectrophotometer using a 1-cm path length (Table 3⇓). The average instrumental differences between microplate and spectrophotometer ranged from 1 to 6% for tannin, 1 to 5% for IRP, <1 to 4% for anthocyanin, and 0.50 to 6.66% for LPP and SPP. The greatest discrepancies were seen in LPP measurements, where the highest percent difference was 6.66% for the 2003 Merlot samples. Mixing by pipetting, as compared to vortex mixing, was found to improve reproducibility within sample sets. The exception was the standard curve, which yielded consistent results with either mixing method.
Summary of instrumental difference between microplate reader and spectrophotometer for phenolic components of 40 commercial wines, including Pinot noir, Merlot, Syrah blends, and Cabernet Sauvignon blends.
Results were within an acceptable range, as this level of variation was occasionally seen in LPP measurements within the same sample set on both instruments. We also compared the reproducibility of the assay on the two instruments by calculating range, standard deviation, and percent coefficient of variation for each phenolic component in a 2003 Merlot sample, using 10 replicates (Table 4⇓). These statistical calculations show that assay values collected using the microplate reader are at least as precise as those collected using the spectrophotometer with a 1-cm path length. Standard deviations for IRP, anthocyanin, LPP, and SPP collected on the microplate reader were actually lower than those collected on the spectrophotometer, with LPP measurements displaying the most significant improvement in precision when measured on the microplate reader as compared with the spectrophotometer.
Comparison of reproducibility on plate reader versus spectrophotometer using 10 replicates of a 2003 Merlot sample (n = 10) assayed for all phenolic components.
Conclusions
A significant improvement in the efficiency of the comprehensive red wine phenolics assay described by Harbertson and colleagues has been demonstrated using a 96-well microplate reader as compared to a traditional spectrophotometer. Both instruments show similar deviations when measuring phenolic material in wines using the comprehensive assay, but this work shows that a microplate reader can be used to measure the phenolic material in 36 wine samples, for example, over 2 to 3 hr with at least as much precision as that of 12 samples assayed on a traditional spectrophotometer over a 3-hr duration. Because the reaction volume has been scaled down to one-fifth the volume, the cost of reagents is also reduced by that same factor. The most significant cost savings was from the decrease in time spent by a laboratory technician to run the assay, with additional savings from the elimination of cuvette disposal and a reduction in the number of 1.5-mL microfuge tubes used and disposed of during the assay. For example, in order to run 36 wine samples in duplicate on a traditional spectrophotometer, it would take 180 disposable cuvettes and 144 microfuge tubes, whereas it would take a maximum of five microplates, which can be reused up to five times, and 72 microfuge tubes to run the same number of samples on a microplate reader.
For winemakers and enologists, the increased efficiency of the assay on a microplate reader would be the strongest argument in favor of using the microplate reader instead of a traditional spectrophotometer since critical fermentation, maceration, and pressing decisions need to be made quickly and confidently during harvest. With an efficient in-house method for phenolic measurement, swift and reliable winemaking decisions can be made on the same day that the sample is submitted to lab for analysis.
- Received May 2006.
- Revision received July 2006.
- Copyright © 2006 by the American Society for Enology and Viticulture
Vol 57 Issue 4
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
