Article pubs.acs.org/JAFC
Relationship between Menthiafolic Acid and Wine Lactone in Wine Joanne Giaccio,† Chris D. Curtin,‡ Mark A. Sefton,† and Dennis K. Taylor*,† †
Downloaded by UNIV OF SHEFFIELD on September 9, 2015 | http://pubs.acs.org Publication Date (Web): September 9, 2015 | doi: 10.1021/acs.jafc.5b03147
Department of Wine Science, School of Agriculture, Food and Wine, The University of Adelaide, Waite Campus, PMB 1, Glen Osmond, South Australia 5064, Australia ‡ The Australian Wine Research Institute, P.O. Box 197, Glen Osmond, South Australia 5064, Australia ABSTRACT: Menthiafolic acid (6-hydroxy-2,6-dimethylocta-2,7-dienoic acid, 2a) was quantified by GC-MS in 28 white wines, 4 Shiraz wines, and for the first time in 6 white grape juice samples. Menthiafolic acid was detected in all but one of the wine samples at concentrations ranging from 26 to 342 μg/L and in the juice samples from 16 to 236 μg/L. Various model fermentation experiments showed that some menthiafolic acid in wine could be generated from the grape-derived menthiafolic acid glucose ester (2b) during alcoholic and malolactic fermentation. Samples containing high concentrations of menthiafolic acid were also analyzed by enantioselective GC-MS and were shown to contain this compound in predominantly the (S)configuration. Enantioselective analysis of wine lactone (1) in one of these samples, a four-year-old Chardonnay wine showed, for the first time, the presence of the 3R,3aR,7aS isomer of wine lactone (1b), which is the enantiomer of the form previously reported as the sole isomer present in young wine samples. The weakly odorous 3R,3aR,7aS 1b form comprised 69% of the total wine lactone in the sample. On the basis of the enantioselectivity of the hydrolytic conversion of menthiafolic acid to wine lactone at pH 3.0 determined previously and the relative proportions of (R)- and (S)-menthiafolic acid in the Chardonnay wine, the predicted ratio of wine lactone enantiomers that would be formed from hydrolysis at ambient temperature of the menthiafolic acid present in this wine was close to the ratio measured, which was consistent with menthiafolic acid being the major or sole precursor to wine lactone in this sample. KEYWORDS: wine lactone, wine, precursor, hydrolysis, enantioselective analysis, aroma, flavor
■
hydrolyzates of grape extracts, Figure 1.12,13 Although both 2a and 2b could be converted to wine lactone in model wine media under forcing conditions, only the acid 2a formed wine lactone at a measurable concentration at ambient temperature and then in low yield only.14 It is possible, however, that the grape-derived glucose ester 2b could act as an indirect precursor to wine lactone if it is first hydrolyzed to acid 2a during winemaking. Enantioselective analysis showed that isomer 1a (previously identified in wine)1,2 and its enantiomer 1b (not yet identified in wine) were formed in near equal amounts from the 6R enantiomer of acid 2a at elevated temperature (45 °C)14 and that the ratio of 1a/1b was variable under simultaneous distillation extraction conditions,14−16 but at ambient temperature, the 1a enantiomer predominated, especially at higher pH.14 No diastereoisomers of wine lactone other than 1a/1b have been observed as hydrolysis products of 2a.14−16 Although the feasibility of wine lactone formation from menthiafolic acid 2a under wine storage conditions has been demonstrated, clarification of the status of this acid as an important precursor of wine lactone isomer 1a requires quantification and enantioselective analysis of 2a in wine. In a preliminary communication, we described the development of a GC-MS method for quantifying menthiafolic acid (2a) and measured the concentration of 2a in several young white wines.17 We describe here the further application of this
INTRODUCTION The potent odorant known as “wine lactone” (1) has been identified among the volatile constituents of white wine1,2 as well as of other products3−5 and has been implicated as a contributor to wine aroma. Hitherto, only one (1a, Figure 1)
Figure 1. Structures of wine lactone (1), wine lactone enantiomers (1a and 1b), menthiafolic acid (2a), and menthiafolic acid glucose ester (2b).
among eight possible stereoisomers of wine lactone had been observed as a wine component,6 and coincidently, this isomer also has the lowest orthonasal detection threshold of these eight isomers.2 The observation that the concentration of wine lactone in a Gewürztraminer wine increased during bottle aging7 led to the hypothesis that this compound might be formed by acid-catalyzed transformation of one or more grapederived precursors.8 Monoterpene acid 2a (sometimes known as menthiafolic acid)9 and the corresponding glucose ester 2b have both been reported as constituents of a Riesling wine,10,11 and acid 2a has been tentatively identified as a component of enzyme © XXXX American Chemical Society
Received: June 25, 2015 Revised: August 26, 2015 Accepted: August 29, 2015
A
DOI: 10.1021/acs.jafc.5b03147 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
out using pentane/ether (95:5, 200 mL, collected as fraction 1; 90:10, 200 mL, fraction 2; 70:30, 100 mL, fraction 3; 70:30 100 mL, fraction 4; 60:40, 200 mL, fraction 5; and 0:100, 200 mL, fraction 6). All fractions were concentrated to 5 mL as above and further concentrated to 200 μL under a stream of nitrogen. Samples were then transferred into a 2 mL GC/MS vial fitted with a 250 μL insert ready for analysis by GC-MS. Wine lactone was detected in fractions 3 and 4. Method Validation for Quantitative Analysis of Menthiafolic Acid in Wine and Grape Juice. Calibration curves for menthiafolic acid (2a) were obtained with spiked standard additions to model wine (pH 3.3), white wine, and grape juice. The analyte (2a) was added to afford the following concentrations: in model wine, 0, 5, 10, 50, 100, 250, 500, and 2000 μg/L; in white wine, 0, 5, 10, 50, 100, 150, 250, 500, 1000, and 2000 μg/L; and in grape juice, 0, 5, 10, 150, 250, 500, and 2000 μg/L. All spiked samples were prepared, extracted, and analyzed in duplicate as previously described.17 For the white wine, six replicates of the 10 and 500 μg/L spiked wine samples were prepared, extracted, and analyzed to test the repeatability of the method. The calibration curves were linear throughout the concentration range, with correlation coefficients (r2) for 2a of 0.9998, 0.9987, and 0.9993 for the model wine, white wine, and grape juice, respectively. For the white wine, the standard deviations were 6.5 and 3.5% at 10 and 500 μg/L, respectively. The slopes of the calibration curves were virtually identical. The LOD was 5 μg/L, and LOQ was 15 μg/L on the basis of visual evaluation. On the day of each set of analyses of the commercial wine and juice samples and of the various ferments described below, a minicurve was prepared in model wine. Three concentration points were prepared in duplicate at 5, 500, and 2000 μg/L and extracted simultaneously with the wine samples. QC’s with additions of 10 and 500 μg/L were also prepared in duplicate and analyzed along with the wine, juice, and fermentation samples, each of which were also analyzed in duplicate. Fermentation Studies with Menthiafolic Acid Glucose Ester (2b). All fermentation treatments were conducted in triplicate with either CDGJM (150 mL) or synthetic wine (150 mL) containing glucose ester 2b at a concentration of 1.32 mg/L. Triplicate control experiments were carried out (a) with inoculated medium (150 mL) containing no glucose ester 2b and (b) with noninoculated medium (150 mL) spiked with glucose ester 2b and stored for the same time and temperature as the ferments. Starter cultures were prepared as follows: S. cerevisiae strain AWRI 838 and Dekkera bruxellensis strain AWRI 1499 were both obtained from the Australian Wine Research Institute culture collection maintained on yeast malt (YM) medium (Amyl Media, Dandenong Australia) supplemented with 1.5% agar and stored at 5 °C. Starter cultures were prepared by adding a loopful of yeast cells to YM (10 mL) and incubating at 28 °C with shaking for 48 h (Saccharomyces) or 72 h (Dekkera). Preadaptation of the cells to the fermentation medium was carried out by inoculating 600 μL (Saccharomyces) or 1 mL (Dekkera) of the starter culture into a 50:50 solution of CDGJM/YM (30 mL) and incubating at 28 °C overnight (Saccharomyces) or for 48 h (Dekkera). Oenococcus oeni (Lallemand VP-41) was reconstituted from a commercial preparation according to manufacturers instructions by addition of the freeze-dried bacteria (0.15 g) to sterilized water (30 mL) at 20 °C, which was then allowed to stand for 15 min with occasional stirring. Alcoholic Fermentations. S. cerevisiae strain AWRI 838 (300 μL of a ca. 2 × 108 cell count culture) in 1:1 YM/CDGJM was inoculated into CDGJM (150 mL) containing glucose ester 2b. Fermentations were incubated at 22 °C for 9 days and were shaken constantly at 120 rpm. Fermentation progress was monitored by weight, and ferments were deemed complete when no further weight loss was recorded. This was verified by analysis of residual glucose and fructose, which showed an average final total sugar content of 6.4 g/L. Malolactic Fermentations. O. oeni (Lallemand VP-41) in sterile water (300 μL) was inoculated into synthetic wine (150 mL), containing the glucose ester 2b, that had previously been stored in an anaerobic chamber to remove all oxygen. The fermentations were incubated under anaerobic conditions for 13 days at 22 °C. Fermentation progress was monitored by malic acid content, which was an average of 0.48 g/L at the end of fermentation.
method to the quantitative and enantioselective analysis of 2a in young and old white and red wines as well as some juice samples. Studies of the fate of glucose ester 2b during fermentation were also conducted to determine whether menthiafolic acid (2a) could be formed from 2b during vinification. Finally, the chirality of wine lactone (1) and menthiafolic acid (2a) in an older wine was also investigated to see whether the isomeric distribution of 1a and 1b in the wine was that expected if 1 was formed primarily from 2a.
Downloaded by UNIV OF SHEFFIELD on September 9, 2015 | http://pubs.acs.org Publication Date (Web): September 9, 2015 | doi: 10.1021/acs.jafc.5b03147
■
MATERIALS AND METHODS
General. All reagents used were purchased from Sigma-Aldrich (Australia). Solvents for GC/MS were CHROMASOLV Plus (SigmaAldrich, Australia) or Pest-grade, purchased from Asis Scientific (Hindmarsh, South Australia), and dispensed using a Puresolv solvent purification system (Innovative Technologies, Massachusetts, USA). Other solvents used were OmniSolv HPLC-grade from VWR (Murrarie, QLD, Australia), with the exception of ethanol that was fractionally distilled food-grade ethanol. The water used was purified by a Milli-Q system. Racemic and (6R)-(E)-2,6-dimethyl-6-hydroxyocta-2,7-dienoic acid (menthiafolic acid, 2a), menthiafolic acid glucose ester 2b, racemic d5-menthiafolic acid, racemic and (3S,3aS,7aR)-3a,4,5,7a-tetrahydro-3,6-dimethylbenzofuran-2-one (wine lactone) were synthesized as described previously.8,14,17 Commercial wine samples were bottled products obtained from local retail outlets. Juice samples were prepared by homogenizing grape samples (ca. 500 g), centrifuging them at 4 °C at 10 000 rpm (17 666 g) for 30 min, and taking the supernatant to be used for further analyses. The pH of Maggie Beer verjuice (a commercial juice product) was adjusted to 3.2 with potassium hydroxide (aqueous, 1 M). Model wine was redistilled ethanol in Milli-Q water (10 + 90, v/ v), saturated with potassium hydrogen tartrate and pH-adjusted to pH 3.2 with L-(+)-tartaric acid (10% w/v in water). Chemically defined grape juice medium (CDGJM) was prepared as described by Ugliano et al.18 The prepared CDGJM was sterilized by filtration with a 0.22 μm membrane. Synthetic wine was produced by fermentation of CDGJM (3 L) which was inoculated with 4 × 106 cells/mL of Saccharomyces cerevisiae AWRI 838 cells preadapted in 50:50 YM/ CDGJM (60 mL) as described below and incubated at 22.5 °C for 14 days. Progress of fermentations was assessed by weight, and completion was confirmed when residual glucose and fructose concentration was
