Subscriber access provided by ECU Libraries
Article
Direct and fast analysis of free and sulfite-bound carbonyl compounds in wine by two-dimensional quantitative proton (1H) and carbon (13C) nuclear magnetic resonance spectroscopy (2D q NMR) Maria Nikolantonaki, Prokopios Magiatis, and Andrew L. Waterhouse Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b01682 • Publication Date (Web): 08 Sep 2015 Downloaded from http://pubs.acs.org on September 13, 2015
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Direct and rapid analysis of free and sulfite-bound carbonyl compounds in wine by twodimensional quantitative proton (1H) and carbon (13C) nuclear magnetic resonance spectroscopy (2D q NMR)
Maria Nikolantonaki,1,3 Prokopios Magiatis,2 and Andrew L. Waterhouse 1*
1
Department of Viticulture and Enology, University of California, Davis, California
95616, United States 2
Department of Pharmacognosy and Natural Products Chemistry, Faculty of Pharmacy,
University of Athens, Panepistimioupolis Zografou 15 771, Athens, Greece 3
Current address: Institut Universitaire de la Vigne et du Vin, Jules Guyot, UMR A
02.102 PAM AgroSup Dijon/Université de Bourgogne, Rue Claude Ladrey, BP 27877, 21078 Dijon Cedex, France
*
To whom correspondence should be addressed
Telephone: +1-530-752-4777 Fax: +1-530-752-0382 E-mail:
[email protected]
1
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
Abstract
2
Recent developments that have accelerated 2D NMR methods and improved quantitation
3
have made these methods accessible analytical procedures, and the large signal dispersion allows
4
for the analysis of complex samples. Few natural samples are as complex as wine, so the
5
application to challenges in wine analysis look promising. The analysis of carbonyl compounds
6
in wine, key oxidation products, is complicated by a multitude of kinetically reversible adducts,
7
such as acetals and sulfonates, so that sample preparation steps can generate complex
8
interferences. These challenges could be overcome if the compounds could be quantified in situ.
9
Here, two-dimensional (1H-1H) homonuclear and heteronuclear (13C-1H) single quantum
10
correlations (COSY and HSQC) nuclear magnetic resonance spectra of undiluted wine samples
11
were observed at natural abundance. These techniques achieve simultaneous direct identification
12
and quantitation of acetaldehyde, pyruvic acid, acetoin, methylglyoxal and α-ketoglutaric acid, in
13
wine with only a small addition of D2O. It was also possible to observe and sometimes quantify,
14
the sulfite, hydrate, and acetal forms of the carbonyl compounds. The accuracy of the method
15
was tested in wine samples by spiking with a mixture of all analytes at different concentrations.
16
The method was applied to 15 wine samples of various vintages and grape varieties. The
17
application of this method could provide a powerful tool to better understand the development,
18
evolution, and perception of wine oxidation, and insight into the impact of these sulfite bound
19
carbonyls on antimicrobial and antioxidant action by SO2.
20
2
ACS Paragon Plus Environment
Page 2 of 30
Page 3 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
21
Introduction
22
Wine-relevant carbonyl compounds, formed as natural fermentation or chemical oxidation
23
products, including acetaldehyde (1), acetoin (2), α-ketoglutaric acid (3), pyruvic acid (4), and
24
methylglyoxal (5), have been proposed as markers of wine aging. The concentration of these
25
substances is tied to winemaking and storage conditions, varies significantly,1 and is especially
26
high in sweet wines made from Botrytis cinerea infected grapes probably due to sugar oxidation.2
27
Analyzing carbonyl compounds in such a complex matrix as wine is very daunting
28
because of their chemical reactivity and instability. First, carbonyls are known to take part in the
29
formation of stable colored, structures arising from the reaction between anthocyanins and 1,
30
glyoxylic acid, and 4.3-5 These pigments have strong bonds between the carbonyl and flavonoid
31
moiety, and are slow to release the carbonyl unit for potential analysis, but one method has been
32
described.6 Second, aldehydes and ketones also react with alcohols to form acetals.7, 8 Some of
33
the nitrogen and sulfur containing compounds, such as the amino acids, proteins and amines can
34
also react with carbonyls by nucleophilic attack and form adducts.9 In these reactions, it is not
35
always clear if the formation of these products is reversible under wine-like conditions. In
36
addition, in grape juice and wine, carbonyl compounds react with sulfur dioxide yielding
37
hydroxysulfonates.10, 11 The sulfite reactions are reversible and dissociation constants for sulfur-
38
bound carbonyl compounds like 1, 3, L-xylosone, 5-keto-D-fluctose, 4 and galacturonic acid
39
have been determined under wine-like acidic conditions by Burroughs and Sparks.10-12
40
Free sulfur dioxide is a key to protecting wine from microbes and oxidation, but an
41
accurate measure of free SO2 is complicated due to these many binding reactions. All of the
42
carbonyl compounds listed above can bind to sulfur dioxide, but the amount depends on their
43
concentrations and dissociation constants of each. And the converse is also true: the influence on 3
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 30
44
the sulfite binding on the concentration of free carbonyl compounds as measured by many
45
conventional methods is poorly described, except in a few cases.13 It would be most useful to
46
have a method that does not initiate the dissociation of compounds, particularly where the
47
dissociation constants and rates are unknown or poorly described.
48
Conventional quantitation methods of free carbonyl compounds in wine include non-
49
specific methods such as distillation or reaction with bisulfite, low sensitivity methods based on
50
colorimetric procedures, and single-compound quantitative enzymatic methods.11, 14 The former
51
methods do not provide compound specific data, and applying the latter methods to
52
simultaneously survey all major wine carbonyls would be time consuming. To date, the most
53
effective alternatives for the analysis of carbonyl compounds are GC-NPD/MS and HPLC-UV. In
54
both cases, sample preparation yields derivatives, such as the o-(2,3,4,5,6-pentafluorobenzyl)-
55
hydroxylamine2, a cysteamine (2-aminoethanthiol)15 or the 2,4-dinitrophenylhydrazone13,
56
derivatives before analysis. The effects of the derivatization steps on the various carbonyl
57
equilibria are known for a few specific equilibria, but otherwise are poorly documented.
16-18
58
Current quantitation methods of bound forms include a sulfonate hydrolysis step under
59
strong alkaline conditions (pH = 11) followed by derivatization after acidification14, 19 or direct
60
acid hydrolysis.13 Thus, levels of bound SO2-bound carbonyl compounds are calculated by the
61
difference between two results: a free carbonyl level measured without hydrolysis and a total
62
carbonyls result that includes hydrolysis.18 However, the derivatization conditions may promote
63
the oxidation of the sample, resulting in the formation of de novo interfering carbonyl
64
compounds.20, 21 In addition the application of the preparation steps may well release carbonyls
65
from the bound forms. Ethylidine bridged flavonoids are reported to release acetaldehyde in acid
66
at 50°C in 20 minutes, and acetals can be hydrolyzed under milder conditions.22 Thus, a method
4
ACS Paragon Plus Environment
Page 5 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
67
that can simultaneously quantitate free and sulfonate-bound carbonyl compounds without
68
hydrolysis could provide insight into the chemical reactivity of these compounds, and could
69
possibly be extended to other carbonyl products.
70
While 2D NMR has been widely used for assignments and identification of various
71
compounds, it has only recently been utilized for the quantitation of metabolites in biological
72
samples.23, 24 This development has been facilitated by accelerated data acquisition and greater
73
instrument stability that has improved quantitative reproducibility.25
74
Our effort was aimed at investigating NMR spectroscopy as a tool for the direct analysis
75
of free and sulfite-bound carbonyls in wine. Its implementation should eliminate the sample
76
preparation steps and resulting artifacts present in existing analytical methods. 1H NMR spectra
77
of wine contains thousands of individual resonances of varying intensities, and dispersion of
78
those signals into two-dimensions may provide an opportunity to measure some signals
79
quantitatively. Here we have evaluated the use of 2D 1H-13C HSQC and 1H-1H COSY spectra for
80
the analysis of multiple free and sulfite-bound carbonyl compounds directly in wine.
81 82
EXPERIMENTAL SECTION
83
Chemicals. Acetaldehyde, 1, pyruvate, 4, acetoin, 2, methylglyoxal, 5, α-ketoglutarate, 3,
84
pyromellitic acid, L-tartaric acid and deuterium oxide were purchased from Sigma-Aldrich, Inc.
85
(St. Louis, MO). Sodium bisulfite, and ethanol were purchased from Acros Organics (Morris
86
Plains, NJ). Water was purified using a Milli-Q system (Millipore, Billerica, MA). All chemicals
87
were of analytical grade or of the highest available purity.
5
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
88
NMR Spectroscopy Experiments. NMR experiments were performed on a Bruker 600
89
MHz spectrometer using a cryoprobe at 289 K. 2D acquisition parameters for the 1H–13C HSQC
90
experiment were as follows: The hsqcetgpsisp2.2 pulseprog was used, with a spectral width of
91
8417 Hz in the F2 dimension and 24147 Hz in the F1 dimension, acquisition time was 0.061 s
92
(F2) and 0.005 (F1), relaxation delay was 1.5 s, 1K (t2) x 256 (t1) data points, 8 scans per t1
93
increment. Processing: Zero filling and FT to 2K x 1K data points after multiplication with qsine
94
filter using SSB=2.2D. Acquisition parameters for the COSY experiment were as follows: The
95
cosygpprqfpulseprog was used, with a spectral width of 9615 Hz, acquisition time was 0.10 s,
96
relaxation delay was 1.5 s, 2K(t2) x 256 (t1) data points, 8 scans per t1 increment. Processing:
97
Zero filling and FT to1K x 1K data points after multiplication with sine filter. All NMR data
98
were processed using Topspin 2.1 NMR (Bruker) software.
99
To test the accuracy of the 2D method for wine samples, spiking experiments were
100
performed in a wine sample with the addition of increasing concentrations 10 to 200 mg L-1 of 2,
101
3, 4, 5 and 1 to 50 mg L-1of 1. A correlation analysis was made between the observed and the
102
added concentration values.
103
NMR description of studied compounds.
104
Acetaldehyde (1): 1H NMR (600 MHz, H2O+D2O) δ 9.665 (1H, q, 2.3 Hz, H-1), 2.225
105 106 107
(3H, d, 2.3 Hz, H-2). 13C NMR (151 MHz, H2O+D2O) δ 207.19 (C-1), 30.74 (C-2) Acetaldehyde hydrated (6): 1H NMR (600 MHz, H2O+D2O) δ 5.235 (1H, q, 5.16 Hz, H1), 1.312 (3H, d, 5.16 Hz, H-2). 13C NMR (151 MHz, H2O+D2O) δ 88.80 (C-1), 23.85 (C-2)
6
ACS Paragon Plus Environment
Page 6 of 30
Page 7 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
108
Acetaldehyde (ethyl hemiacetal)(7): 1H NMR (600 MHz, H2O+D2O) δ 4.94 (1H, q, 5.2
109
Hz, H-1), 3.77 (1H, dt, Ha-CH2). 3.51 (1H, dt, Hb-CH2), 1.17 (3H, CH3), 1.303 (3H, d, 5.2 Hz,
110
H-2). 13C NMR (151 MHz, H2O+D2O) δ 94.69 (C-1), 63.31 (CH2), 22.62 (C-2), 14.87 (CH3)
111
Acetaldehyde-bisulfite adduct (8): 1H NMR (600 MHz, H2O+D2O) δ 4.54 (1H, q, 6.45
112
Hz, H-1, overlapped), 1.46 (3H, d, 6.45 Hz, H-2). 13C NMR (151 MHz, H2O+D2O) δ 80.99 (C-
113
1), 17.47 (C-2)
114
Acetoin (2): 1H NMR (600 MHz, H2O+D2O) δ 4.40 (1H, q, 7.15 Hz, H-3), 2.21 (3H, s,
115
H-1), 1.365 (3H, d, 7.15 Hz, H-4). 13C NMR (151 MHz, H2O+D2O) δ 215.94 (C-2), 73.60 (C-3),
116
25.55 (C-1), 18.90 (C-4)
117
Acetoin-bisulfiteadduct (isomer I) (9a,b): 1H NMR (600 MHz, H2O+D2O) δ 4.165 (1H,
118
q, 6.54 Hz, H-3), 1.43 (3H, s, H-1), 1.24 (3H, d, 6.54 Hz, H-4). 13C NMR (151 MHz, H2O+D2O)
119
δ 89.75 (C-2), 70.06 (C-3), 17.87 (C-4), 17.26 (C-1)
120
Acetoin-bisulfiteadduct (isomer II) (9c,d): 1H NMR (600 MHz, H2O+D2O) δ 4.165 (1H,
121
q, 6.54 Hz, H-3), 1.48 (3H, s, H-1), 1.24 (3H, d, 6.54 Hz, H-4). 13C NMR (151 MHz, H2O+D2O)
122
δ 89.62 (C-2), 69.98 (C-3), 16.98 (C-4), 16.81 (C-1)
123 124 125 126
Ketoglutaric acid (hydrated) (10a): 2.43 (2H, t, 7.5 Hz, CH2), 2.11 (2H, t, 7.5 Hz, CH2). 13
C NMR (151 MHz, H2O+D2O) δ 29.29, 34.08 Ketoglutaric acid (cyclic) (10b): 2.98 (2H, br, CH2), 2.66 (2H, t, 6.7 Hz, CH2). 13C NMR
(151 MHz, H2O+D2O) δ 34.56, 28.31
127
Ketoglutaric acid-bisulfite adduct (11b): 1H NMR (600 MHz, H2O+D2O) δ 2.56 (1H, m,
128
CH), 2.49 (1H, m, CH), 2.34 (1H, m, CH), 2.25 (1H, m, CH). 13C NMR (151 MHz, H2O+D2O) δ
129
177.85 (CO), 172.65 (CO), 90.66 (C-2), 29.53, (CH2), 29.12 (CH2).
7
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Pyruvic acid (carbonyl form) (4): 1H NMR (600 MHz, H2O+D2O) δ 2.42 (3H, s, CH3).
130 131
13
C NMR (151 MHz, H2O+D2O) δ 200.16 (C-2), 174.83 (CO), 26.56 (CH3). Pyruvic acid (hydrated form)(12): 1H NMR (600 MHz, H2O+D2O) δ 1.57 (3H, s, CH3).
132 133
13
C NMR (151 MHz, H2O+D2O) δ 175.60 (CO), 93.30 (C-2), 25.77 (CH3). Pyruvic acid-bisulfite adduct (13): 1H NMR (600 MHz, H2O+D2O) δ 1.73 (3H, s, CH3).
134 135
13
C NMR (151 MHz, H2O+D2O) δ 173.52 (CO), 88.56 (C-2), 21.37(CH3). Methylglyoxal hydrated form I(14a): 1H NMR (600 MHz, H2O+D2O) δ 5.26 (1H, s, H-
136 137
3), 2.29 (3H, s, CH3).
138
(CH3).
140
3), 1.36 (3H, s, CH3).
141
(CH3).
143
C NMR (151 MHz, H2O+D2O) δ 209.50 (CO), 90.34 (C-3), 25.01
Methylglyoxal hydrated form II (14b): 1H NMR (600 MHz, H2O+D2O) δ 5.07 (1H, s, H-
139
142
13
13
C NMR (151 MHz, H2O+D2O) δ 95.76 (C-2), 92.55 (C-3), 22.07
Methylglyoxal-bisulfite adduct (15): 1H NMR (600 MHz, H2O+D2O) δ 5.20 (1H, s, H-3), 2.43 (3H, s, CH3). 13C NMR (151 MHz, H2O+D2O) δ 206.15 (CO), 88.59 (C-3), 27.76(CH3).
144
Sample preparation for NMR analysis. To 0.45 ml of wine or standard solution 50µL of
145
a solution containing 0.78 mM pyromellitic acid in D2O was added in a 5 mm NMR tube. The
146
signal from the pyromellitic acid served as an internal standard (IS). After thorough mixing, 0.5
147
mL of each sample was transferred to a 5 mm-o.d Wilmad® NMR tube for analysis.
148
Quantitation. Standard addition procedure. The concentrations of free and sulfite-bound
149
carbonyls were calculated using standard addition procedure. A model mixture for standard
150
additions was prepared with five targeted wine-relevant free carbonyls (1, 2, 3, 4, 5), dissolved in
151
1 ml H2O, each at 1 gm-L-1. This stock solution was added to a model wine (12% ethanol, pH =
8
ACS Paragon Plus Environment
Page 8 of 30
Page 9 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
152
3.5, 5 gL-1 tartaric acid) to obtain a range of concentrations varying between 10 and200 mgL-1 for
153
2, 3, 4, 5, and 1 to 50 mgL-1 for 1. Each standard solution was measured five times as described
154
above.
155
The calibration curves for the sulfite-carbonyl adducts were prepared using 1 gmL-1stock
156
solution as above, diluted into model wine, to obtain concentrations between 10 to 200 mgL-1for
157
all compounds. To five tubes, each containing 500 µL model wine spiked with carbonyls, 25 µL
158
of sulfur dioxide (10 mM) was added. The mixture was mixed for 10s and incubated for 1h to
159
achieve equilibrium. For each sample, the sulfite-adduct was analyzed as described above. It is
160
presumed that all the added carbonyl would be in the sulfite adduct form under these conditions.
161
For each free and sulfite-bound carbonyl compound, a standard addition curve was fitted
162
by the linear regression equation: V = a[c] + b, where V represents the 2D peak volume
163
integration ratio between pyromellitic acid (IS) and each free and/or sulfite-bound carbonyl
164
compound, and [c] the concentration (mgL-1) in wine. Table 1 and Table 2 show the number of
165
signals that were used for compound identification. However, due to overlap of signals even in
166
the 2D spectra, only a subset of these signals could be used for quantitation. The observed
167
concentration of each compound was calculated by a b/a ratio, where a is the slope and b the y-
168
intercept of the linear regression curve.
169
Analytical evaluation. The standard addition curve linearity was evaluated by its
170
regression coefficient r2, Table 3.To test the accuracy of the 2D method for wine samples, spike
171
experiments were performed using a wine sample with the addition of two different
172
concentrations of each compound (15, 50 mgL-1for 2, 3, 4, 5 and 3, 10 mgL-1for 1). The accuracy
173
was expressed as the percent recovery (%R) calculated for each compound using: %R = (Cs-
174
Cu)/Csa, where Cs: the measured concentration in spiked aliquot; Cu: the measured concentration
9
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 30
175
in unspiked aliquot and Csa: the concentration of spike added. The precision of NMR
176
measurement was determined on five successive analysis of the same sample at two
177
concentrations (Table 4) and evaluated via the percentage of the coefficient variation (CV% =
178
SD 100/AV), where SD is the standard deviation and AV the average value of replicates. The
179
limit of detection (LOD) and quantitation (LOQ) of each analyte were estimated based on the
180
standard deviation of the response and the slope (a) of the calibration curve according to the
181
formulas: LOD = 3.3 (SD/a) and LOQ = 10 (SD/a). Because of the reactivity and instability of
182
the carbonyl compounds in wine, method validation was performed using a wine sample
183
containing low levels of targeted compounds.
184
RESULTS AND DISCUSSION
185
Assignments of free and sulfite-bound carbonyl signals. NMR spectra (1D and 2D) of
186
model wine solutions containing 1 mM carbonyl compounds (1, 2, 3, 4, and 5), with and without
187
6 mM (384 mg/L) sulfur dioxide revealed the presence of the free forms, hydrates, diethyl acetals
188
and sulfonate adducts, but not all for every compound. The progress of the conversions was
189
monitored by 1H NMR. In all reaction mixtures, the equilibria were reached within a few
190
minutes.
191
Identification of free and sulfite-bound acetaldehyde. The 1H-NMR of acetaldehyde, 1, in
192
model wine exhibited three sets of peaks; one corresponding to the carbonyl form (1), one
193
corresponding to the hydrate form (6) and one corresponding to the ethyl hemiacetal (7). The
194
expected resonances of 1 consisted of a highly deshielded quartet at 9.66 ppm (aldehydic proton)
195
and a doublet at 2.22 ppm corresponding to the methyl group. The aldehydic carbon was
196
observed at 207.19 ppm. The hydrate 6 revealed a proton at 5.23 ppm (quartet) coupled with the
197
methyl group at 1.31 ppm (doublet). The dihydroxy carbon was observed at 88.8 ppm confirming
10
ACS Paragon Plus Environment
Page 11 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
198
the presence of two oxygen atoms. Compound 7 was a hemiacetal (4.94 ppm/94.69 ppm) also
199
containing an ethyl group identified in the carbon spectrum with two peaks at 63.31 and 14.87
200
ppm. The two protons of the CH2 group were observed at 3.77 and 3.51 ppm as two doublets of
201
triplets. The methyl group overlapped with the ethanol peak. After addition of sodium
202
metabisulfite a new set of peaks appeared, corresponding to the bisulfite adduct 8. The methyl
203
group of the adduct was observed as a doublet at 1.46 ppm while the methine proton overlapped
204
with the peak of tartaric acid at 4.54, as clarified by the COSY spectrum. In the HSQC spectrum
205
the methine proton was correlated with a carbon at 80.99 as expected for a bisulfite adduct.
206
Identification of free and sulfite-bound acetoin. In contrast to acetaldehyde, racemic 2 in
207
model wine was observed as a single compound in the free carbonyl form, with no hydrate or
208
acetals present. The 1H-NMR spectrum showed one methyl adjacent to a carbonyl at 2.21 ppm,
209
one methyl at 1.365 ppm, and one methine on the oxygenated carbon at 4.40 ppm. The carbonyl
210
of 2 was observed at 215.94 ppm in the 13C spectrum. Addition of sodium metabisulfite led to the
211
appearance of two new compounds. Indeed the addition of sulfite group at position 2 of racemic 2
212
leads to 2 diasteriomers, as two pairs of enantiomers (9 a-d). Each diastereomer gave four distinct
213
signals in the carbon spectrum, and in the 1H-NMR two distinct signals for the methyl groups of
214
position 1 (1.43 and 1.48 ppm) and two over-lapping signals for positions 3 and 4 at 4.165 and
215
1.24 ppm respectively.
216
Identification of free and sulfite-bound α-ketoglutaric acid. Ketoglutaric acid was more
217
complex than the others because in addition to the free form (3) and the hydrated form (10a), a
218
cyclic form could also be observed (10b). The equilibrium was dependent on the pH of the
219
solution and also on the concentration of the compound. However at the pH of the model wine,
220
and the concentration range used for the calibration curve, the dominant form was the cyclic 10b,
11
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 30
221
but with some contribution from the free 3 via rapid (on the NMR timescale) interconversion. It
222
was observed as two peaks (one triplet and one distorted broad peak) corresponding to two
223
methylenes. The formation of the SO2 adduct would lead to either 11a or 11b. There were four
224
distinct multiplets corresponding to the four protons of the two methylenes, and two methylenes
225
in the 13C spectrum plus an oxysulfated quaternary carbon and the two carbonyls. A large
226
difference in the chemical shift of the two carbonyls suggests that one correspond to a free
227
carboxylic acid and one to a lactone, leading to the conclusion that the observed form was 11b.
228
Identification of free and sulfite-bound pyruvate. Pyruvic acid in model wine was
229
observed as a mixture of the carbonyl form (4) and the hydrate form (12). Each form gave a
230
distinct peak for the methyl group, 4 at 2.42 ppm and 12 at 1.57 ppm, assigned using HSQC and
231
13C NMR . The carbonyl form was identified by the signal at 200.16 ppm while the hydrate by
232
the oxygenated carbon at 93.3 ppm. As expected, the methyl of the carbonyl form was deshielded
233
compared to the methyl of the hydrated form. The addition of the sulfite group led to the
234
appearance of a new methyl signal at 1.73 ppm corresponding to compound 13.
235
Identification of free and sulfite-bound methylglyoxal. Compound 5 in model wine was
236
observed as a mixture of the two hydrated forms (14a, 14b) while the carbonyl form 5 was not
237
observed. The major hydrate form 14a corresponded to the compound with one hydrated
238
aldehyde while the minor compound 14b corresponded to the case where both carbonyls were
239
hydrated. In the first there was one carbonyl at 209.5 ppm and a doubly oxygenated carbon at
240
90.3 ppm. In the second case the molecule showed two doubly oxygenated carbons at 92.5 and
241
95.7 ppm. The addition of sodium metabisulfite led to only one observable adduct, 15, where the
242
sulfite group had been added to the aldehydic carbon. The adduct showed one carbonyl at 206.1
243
ppm, one oxysulfated methine carbon and one methyl group.
12
ACS Paragon Plus Environment
Page 13 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
244
Quantitative 2D NMR analysis.
245
Calibration of 2D NMR results. Our strategy to quantify free and sulfur-bound carbonyls
246
compounds directly in wine was to apply 2D 1H-1H COSY and 1H-13C HSQC techniques
247
(Figure2) which have not been used previously in wine quantitation. For the quantitative 2D
248
NMR experiments, 10% D2O was added to wine samples for locking.
249
Since these techniques had not been applied to quantitation of wine components, it was
250
important to verify whether volumes of 2D signals could be proportional to the concentration of
251
the corresponding analyte. For quantitation, the non-sulfonate forms of each compound (formed
252
instantly) were summed and expressed in carbonyl form equivalents. So the concentration of 1, 4,
253
and 5 were expressed as the sum of 1,6, plus 7; 4, plus 12;and 14a, plus 14b, respectively.
254
Method Validation. The method showed satisfactory linearity for all free carbonyl
255
compounds within the range of 1.2-10 mgL-1 to 200 mgL-1 for 2, 3, 4, 5 (see Table 3), and 0.7
256
mgL-1 to 50 mgL-1 for 1, with the corresponding regression correlation coefficients of (r2) above
257
0.990. Moreover, calibration curves showed strong linear (r2 > 0.989) responses for the sulfite-
258
bound carbonyls, 8, 13, and 15, within the ranges of 2.5-600,12.3-432 and 8.6-435 mg L-1,
259
respectively. Calibration curves of all free and sulfite-bound carbonyls showed the same order of
260
magnitude for their slope value, suggesting a similar sensitivity of the analytical method for all
261
the assessed compounds. LOQ were observed for 1/8, 2, 3, 4/13 and 5, at 1.1/7, 13.2, 16.8,
262
3.6/10.3, and 12.5/19.1 mg L-1, respectively. However, because of the apparent high dissociation
263
constant, and consequently low sulfite binding power of 2, none of its sulfite adduct, 9abcd,
264
could be detected in wine-like conditions containing a 4-fold excess of sulfur dioxide (relative to
265
the concentration of 2). In addition to that, severe peak overlap in HSQC spectra was encountered
13
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 30
266
on the assigned 11b in wine samples. As a consequence of these problems, neither the 9abcd nor
267
the 11b species could be observed in the course of wine analysis.
268
The accuracy and precision of the method were assessed using low (15 mg L-1 for 2, 3,
269
4/13, 5/15, 8, and 3 mg L-1 for 1) and high (50 mg L-1 for 2, 3, 4/13, 5/15, 8, and 10 mg L-1 for 1)
270
concentrations in triplicate. All samples were assessed over two days and this stability assay
271
showed no significant reduction of analyte concentrations in the wine matrix at either low or high
272
concentrations. The accuracy of the method was calculated as the ratio of the mean observed
273
increased concentration and the known spiked amount added to the wine matrix and expressed as
274
the percent recovery for the three samples (Table 4). Percent recovery (IS-normalized) results in
275
wine for 2, 3, 4/13, 5/15 and 8 were within 92 and 116 %, highlighting the accuracy of the
276
quantitation in wine. The coefficient of variation ranged from 0.5% to 7.2%, demonstrating good
277
precision for an NMR based method.
278
Sensitivity and limitations of the method. The proposed method is based on the fact that
279
cross-peak volumes observed in the 2D 1H-1H COSY and 1H-13C HSQC NMR spectrum can be
280
correlated with the concentration of the analytes. The lower limit of detection of the proposed
281
method will depend upon several experimental parameters which govern cross-peak volumes in
282
the COSY and HSQC spectra. With the present experimental results we have been able to
283
measure the concentration of free and sulfite bound low molecular weight carbonyl compounds
284
directly in wine to within a few tens of milligrams per liter range. In our experiments, performed
285
in a 600 MHz NMR magnet equipped with a cryo-probe, we have used approximately 1h for
286
recording the COSY and HSQC spectra of the wine samples. We have found (data not shown)
287
that by increasing to 9 h the signal averaging recording time the LOD of all tested analytes can be
288
lowered. Nevertheless, considering that the quantitation limits of all analytes were adequately
14
ACS Paragon Plus Environment
Page 15 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
289
below the reported mean levels of the tested carbonyls in wines when they were measured with 1
290
h recording time, these experimental parameters were set as the default in order to avoid
291
excessive analysis time.
292
Method application in wine. Several California wines (n = 15) from different vintages
293
(1974 - 2012) were analyzed for 1/8, 2, 3, 4/13, and 4/15 using the above described method. The
294
observed concentrations and standard deviations are shown in Figure 3. The mean concentrations
295
of 1, 2, 3, 4 and 5 were found to be 17.5 ± 32.6, 12.7 ± 7.8, 106.5 ± 56.7,36 ± 27.8 and 6.4 ± 2.8
296
mg L-1 respectively. The observed concentrations of these carbonyl compounds from 2D COSY
297
and HSQC spectra are well within the range of those reported previously.1,
298
survey of targeted carbonyl-bisulfite adducts, reported for the first time, show distinct differences
299
and variation in their concentration between the 8, 13 and 15 bisulfites. In the wines tested, 8 and
300
13 were found at higher concentrations, 28.2 ± 60.2 and 110 ± 106.2 mgL-1, respectively,
301
followed by 15 (7.8 ± 8.8 mgL-1). Not surprisingly, acetaldehyde and pyruvate bisulfite adducts
302
have some of the lowest reported dissociation constants of quantitatively important carbonyls,11
303
whereas the significance of the methyl glyoxal-bisulfite adducts, 14 and 15 has not been clarified
304
in dry wines. However, the identification of these bisulfites at significant concentrations in wines
305
older than one year (data not shown) indicates stability and apparently slow reaction of the free
306
carbonyl with other substrates under wine aging conditions. In addition, the distribution of these
307
bisulfites in the assessed wines showed no correlation with the vintage or the grape variety,
308
suggesting a that other factors, perhaps fermentation conditions or oxidation during processing
309
contribute to the free and sulfite bound carbonyls levels in wine.
16
Moreover, the
310
Influence of sulfur dioxide concentration at bottling on the levels of free and sulfite
311
bound carbonyls in wines. The development of free and sulfite bound carbonyls during bottle
15
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 30
312
aging in the presence of different levels of sulfur dioxide was studied by the screening of these
313
compounds in a wine (Cabernet sauvignon, Oakville, Napa, 1997) bottled with 30 and 120 mgL-
314
1
315
sulfur dioxide addition generally resulted in both treatments having little accumulation of 1,
316
indicating the control of chemical oxidation even at the low sulfur dioxide concentration (30
317
mgL-1), or a reaction pathway comparable to the reaction rate.
sulfur dioxide (Figure 4). The analysis of the wines at 15 years of bottle aging showed that the
318
High sulfur dioxide containing treatment (120 mg L-1) increased 8, 13 and 15 bisulfite
319
adducts formation by 18%, >100% and >100%, respectively, indicating that hydroxysulfonate
320
adducts have substantial stability under wine aging conditions. Interestingly, in the case of high
321
sulfur dioxide treatment, between 1, 4 and 5, only 4 was found present in free form. This suggests
322
that, under wine aging conditions, 4 is a weaker sulfur dioxide binder compared to 1 and 5.
323
However, sulfur dioxide excess at bottling had not affected 10b concentration in wines,
324
indicating weak sulfite binding by 3. Moreover, 1 showed the highest binding with sulfur dioxide,
325
since even in the low sulfur dioxide treatment (30 mg L-1) only its hydroxysulfonate adduct, 8,
326
was detected. This observation is supportive of the low dissociation constant (Ks) reported for 8
327
in model wine solution.2,
328
resulted in much higher levels of both 5 and its hydroxysulfonate, 15.
12
One mystery is why the elevated level of sulfites during aging
329
CONCLUSIONS
330
2D NMR can be used for the accurate quantitation of carbonyl compounds and their
331
sulfite bound forms in wine. This analytical capacity paves the way toward a number of studies to
332
better understand how wine carbonyls could be formed or react with other wine constituents.
333
Such studies could provide insight into yeast metabolism during fermentation, when a large
334
fraction of the carbonyls are formed, or as a result of oxidation, such as during micro-
16
ACS Paragon Plus Environment
Page 17 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
335
oxygenation treatments. While the hydroxysulfonate of acetaldehyde is often attributed to the
336
“bound” sulfites observed in wine, this data clearly shows there are several additional carbonyls
337
that are found in sulfite-bound forms in substantial amounts, adding up to some tens of
338
milligrams per liter. The data provided here leads to the question of whether or not these
339
hydroxysulfonates can contribute to, or subtract from, microbial stability or antioxidant
340
protection of wines provided by SO2, particularly as the wine ages and sulfites level decline.
341 342
AKNOWLEDGMENT
343
We sincerely thank the American Vineyard Foundation (AVF) for its financial support.
344
17
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
345
Page 18 of 30
References
346
(1)
Jackowetz, J. N.; Mira de Orduña, R. Food Control 2013, 32, 687-692.
347 348
(2)
Barbe, J.-C.; de Revel, G.; Joyeux, A.; Lonvaud-Funel, A.; Bertrand, A. Journal of Agricultural and Food Chemistry 2000, 48, 3413-3419.
349 350
(3)
Bakker, J.; Timberlake, C. F. Journal of Agricultural and Food Chemistry 1997, 45, 3543.
351 352
(4)
Fulcrand, H.; Cheynier, V.; Oszmianski, J.; Moutounet, M. Phytochemistry 1997, 46, 223227.
353 354
(5)
Fulcrand, H.; Benabdeljalil, C.; Rigaud, J.; Cheynier, V.; Moutounet, M. Phytochem. 1998, 47, 1401-1407.
355 356
(6)
Drinkine, J.; Lopes, P.; Kennedy, J. A.; Teissedre, P. L.; Saucier, C. Journal of Agricultural and Food Chemistry 2007, 55, 1109-1116.
357 358
(7)
Lorette, N. B.; Howard, W. L.; Brown, J. H. The Journal of Organic Chemistry 1959, 24, 1731-1733.
359
(8)
Peterson, A. L.; Gambuti, A.; Waterhouse, A. L. Tetrahedron 2015, 71, 3032-3038.
360
(9)
Wlodek, L. Acta Biochim. Pol. 1988, 35, 307-317.
361 362
(10)
Burroughs, L. F.; Sparks, A. H. Journal of the Science of Food and Agriculture 1973, 24, 199-206.
363 364
(11)
Burroughs, L. F.; Sparks, A. H. Journal of the Science of Food and Agriculture 1973, 24, 207-217.
365 366
(12)
Burroughs, L. F.; Sparks, A. H. Journal of the Science of Food and Agriculture 1973, 24, 187-198.
367
(13)
Han, G.; Wang, H.; Webb, M. R.; Waterhouse, A. L. Talanta 2015, 134, 596-602.
368
(14)
Ough, C. S.; Amerine, M. A. 2d ed. Wiley & Sons, New York 1988.
369 370
(15)
Lau, M. N.; Ebeler, J. D.; Ebeler, S. E. American Journal of Enology and Viticulture 1999, 50, 324-333.
371 372
(16)
Elias, R. J.; Laurie, V. F.; Ebeler, S. E.; Wong, J. W.; Waterhouse, A. L. Anal. Chim. Acta 2008, 626, 104-110.
373
(17)
Jackowetz, J. N.; Mira de Orduña, R. Food Chemistry 2013, 139, 100-104.
374 375
(18)
de Azevedo, L. C.; Reis, M. M.; Pereira, G. E.; da Rocha, G. O.; Silva, L. A.; de Andrade, J. B. Journal of Separation Science 2009, 32, 3432-3440.
376 377
(19)
Lea, A. G. H.; Ford, G. D.; Fowler, S. International Journal of Food Science & Technology 2000, 35, 105-112.
378
(20)
Wildenradt, H. L.; Singleton, V. L. Am. J. Enol. Vitic. 1974, 25, 119-126.
379
(21)
Waterhouse, A. L.; Laurie, V. F. Am. J. Enol. Vitic. 2006, 57, 306-313.
380
(22)
Chiang, Y.; Kresge, A. J. Journal of Organic Chemistry 1985, 50, 5038-5040. 18
ACS Paragon Plus Environment
Page 19 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
381 382
(23)
Gronwald, W.; Klein, M. S.; Kaspar, H.; Fagerer, S. R.; Nürnberger, N.; Dettmer, K.; Bertsch, T.; Oefner, P. J. Analytical Chemistry 2008, 80, 9288-9297.
383
(24)
Rai, R. K.; Tripathi, P.; Sinha, N. Analytical Chemistry 2009, 81, 10232-10238.
384 385
(25)
Van, Q. N.; Issaq, H. J.; Jiang, Q. J.; Li, Q. L.; Muschik, G. M.; Waybright, T. J.; Lou, H.; Dean, M.; Uitto, J.; Veenstra, T. D. Journal of Proteome Research 2008, 7, 630-639.
386 387
(26)
Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. The Journal of Organic Chemistry 1997, 62, 7512-7515.
388
19
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 30
Table and Figure captions Table 1:Assignment of 2D 1H-1HCOSY NMR signals of free and sulfite-bound carbonyl compounds in wine. Table 2:Assignment of 2D 1H - 13C HSQC NMR signals of free and sulfite-bound carbonyl compounds in model wine and real wine that were used for calibration and quantitation. Table 3:Calibration line parameters average concentrations, standard deviations, and ranges of selected wine free and sulfite bind carbonyls (n = 15). The compound numbering is defined in Figure 1. *1; 4; 5 concentrations are expressed are the sum of 1,6,7; 4, 12; 14a, 14b,respectively. Table 4: Method accuracy and precision estimated by performing triplicate analyses of a wine sampled spiked with two known concentrations of each free and sulfite bound carbonyl compound. The compound numbering is defined in Figure 1. *1; 4; 5 concentrations are expressed as the sum of 1,6,7; 4, 12; 14a, 14b,respectively. Figure 1: Structures of identified free and sulfite bound carbonyl compounds. Acetaldehyde (1), hydrated acetaldehyde (6), acetaldehyde ethyl hemiacetal (7), acetaldehyde bisulfide (8), acetoin (2), acetoin bisulfite (9ab, isomer I), acetoin bisulfite (9cd, isomer II), αketoglutaric acid (3), hydrated ketoglutaric acid (10a), hydrated ketoglutaric acid (10b, cyclic form), α-ketoglutaric acid bisulfite (11a), α-ketoglutaric acid bisulfite (11b, cyclic form), pyruvic acid (4), hydrated pyruvic acid (12), pyruvic acid bisulfite (13), methylglyoxal (5), hydrated methylglyoxal (14a, form I), hydrated methylglyoxal (14b, form II), and methylglyoxal bisulfite (15).
20
ACS Paragon Plus Environment
Page 21 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figure 2: Portion of the (A) 1H-13C HSQC and (B) 1H-1HCOSY 600 MHz cryo-probe spectrum of identified free and sulfite bind carbonyl compounds in wine sample. The compound numbering is defined in Figure 1. Figure 3: Average values along with the standard deviation of the concentration of free and sulfite bound carbonyl compounds of wine samples (n=15). The compound numbering is defined in Figure 1. *1; 4; 5 concentrations are expressed as the sum of 1,6,7; 4, 12; 14a, 14b respectively. Figure 4: Quantitation of free and sulfite bind carbonyls compounds in wine samples (Cabernet Sauvignon from Oakville, Napa, 199 9, spiked at bottling with 20 and 120 mg/L of sulfites. All quantitation assays were performed in April 2013.The compound numbering is defined in Figure 1. *1; 4; 5 concentrations are expressed as the sum of 1,6,7; 4, 12; 14a, 14b,respectively.
21
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Table 1
compound 1 2 6 7 8 9ab 9cd
chemical shift (δ) (ppm) 1 1 H H assignment 9.67 2.23 H1-H2 1.36 4.40 H3-H4 5.24 1.31 H1-H2 4.94 1.30 H1-H2 4.54 1.46 H1-H2 1.24 4.16 H3-H4 1.24 4.00 H3-H4
22
ACS Paragon Plus Environment
Page 22 of 30
Page 23 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Table 2 chemical shift (δ) (ppm) 1 13 compound H C assignment 2.21 25.53 H1/C1 2 2.37 26.90 H3/C3 4 4.16 70.06/69.98 H3/C3 9abcd 2.68 28.24 CH2 10b 2.56 29.53 CH2 11b 2.49 29.53 CH2 2.34 29.12 CH2 2.25 29.12 CH2 1.56 25.81 H3/C3 12 1.72 21.28 H3/C3 13 2.29 25.18 CH3 14a 1.35 22.15 CH3 14b 2.42 27.73 CH3 15
NOTE The spectra were calibrated according to the ethanol peak at 1.17 ppm (proton) and 17.47 ppm (carbon) according to 26.
23
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 24 of 30
Table 3
compounds calibration curve fitted regression equation
r2
linear range (mgL-1)
validation range (mgL-1)
wine conc (mgL-1) mean ± SD
1*
y = 0.0022x + 0.008
0.997
0.7-50
1.1-50
17.5 ± 32.6
2
y = 0.0057x + 0.007
0.992
9.8-200
13.2-200
12.7 ± 7.8
4*
y = 0.0039x + 0.006
0.997
1.2-200
3.6-200
36.0 ± 27.8
5*
y = 0.0032x + 0.004
0.994
8.6-150
12.5-200
6.4 ± 2.8
8
y = 0.0052x + 0.080
0.989
2.5-600
7.0-600
28.2 ± 60.2
10b
y = 0.0045x + 0.022
0.990
10.3-200
16.8-200
106.5 ± 56.7
13
y = 0.0034x + 0.062
0.998
8.2-432
10.3-432
110.0 ± 106.2
15
y = 0.0037x + 0.001
0.994
16.2-435
19.1-435
7.8 ± 8.8
24
ACS Paragon Plus Environment
Page 25 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Table 4
compounds
1* 2 4* 5* 8 10b 13 15 1
concentration accuracy (%R)1
precision SD
CV (%)
3
102 ± 0.01
0.00
2.5
10
98 ± 0.01
0.00
2.9
15
100 ± 0.02
0.32
7.1
50
92 ± 0.01
0.02
4.2
15
102 ± 0.10
0.06
3.4
50
99 ± 0.02
0.01
2.7
15
98 ± 0.01
0.07
4.3
50
109 ± 0.06
0.01
3.6
15
92 ± 0.07
0.06
1.2
50
96 ± 0.02
0.01
0.8
15
84 ± 0.25
0.14
5.6
50
116 ± 0.38
0.08
6.1
15
97 ± 0.01
0.21
7.2
50
113 ± 0.04
0.13
5.4
15
96 ± 0.04
0.15
4.5
50
98 ± 0.03
0.07
3.5
%R = percentage recovery
25
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 1
26
ACS Paragon Plus Environment
Page 26 of 30
Page 27 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figure 2
27
ACS Paragon Plus Environment
Analytical Chemistry
Figure3
40
120
35
100
30
80
25
60
20
40
15
20
10
0
5
-20
Mean Mean±SE Mean±SD
-40
1*
2
4*
13
5*
15
28
ACS Paragon Plus Environment
1
0
n
Conentration (mg/L)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 28 of 30
8
10b
Page 29 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figure 4
Wine-low SO2 Wine-High SO2
1*#
8#
2#
4*#
13#
29
ACS Paragon Plus Environment
5*#
15#
10b"
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
2D qNMR
HSO3-
ACS Paragon Plus Environment
Page 30 of 30