Structural changes of sinapic acid and sinapine bisulfate during autoclaving with respect to the development of colored substances

Article abstract of DOI:10.1007/s11746-999-0021-7

Structural changes in sinapic acid during autoclaving were studied using spectral analysis, thin-layer chromatography, high-performance liquid chromatography, nuclear magnetic resonance (NMR), and mass spectroscopy. Color properties of sinapic acid and its derivatives were studied by determining the transmittance spectrum, calculating the Commission Internationale de l'Eclairage 1931 tristimulus values and converting to Hunter L a b values. It was found that the colorless sinapic acid aqueous solution (100 μg/mL) turned yellow after 15 min in an autoclave at 121 °C and 0.1 MPa. Filtering the yellow aqueous solution through a 0.45-μm filter removed a brown solid consisting of at least three undetermined colored substances and left a yellow liquid. A newly developed yellow substance, syringaldehyde, was identified in the liquid phase by comparing the NMR and mass spectra of the unknown with those of authentic syringaldehyde. Thomasidioic acid was also found in the liquid phase. Under the same autoclaving conditions, sinapine bisulfate showed no evidence of any structural or color changes.

Full text of DOI:10.1007/s11746-999-0021-7

Structural Changes of Sinapic Acid and Sinapine Bisulfate During
Autoclaving with Respect to the Development
of Colored Substances
a
a,
b
R. Cai , S.D. Arntfield *, and J.L. Charlton
a
b
Departments of Food Science and Chemistry, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2
ABSTRACT: Structural changes in sinapic acid during auto- (2,13,14). The severe moist-heat treatment, in the cooker and
claving were studied using spectral analysis, thin-layer chroma- desolventizer unit, darkens the meal color and denatures the
tography, high-performance liquid chromatography, nuclear
magnetic resonance (NMR), and mass spectroscopy. Color
properties of sinapic acid and its derivatives were studied by
determining the transmittance spectrum, calculating the Com-
mission Internationale de l’Eclairage 1931 tristimulus values
and converting to Hunter L a b values. It was found that the col-
orless sinapic acid aqueous solution (100 µg/mL) turned yellow
after 15 min in an autoclave at 121°C and 0.1 MPa. Filtering
the yellow aqueous solution through a 0.45-µm filter removed
protein (11). The resulting commercial canola meals are less
than ideal for food use partly because of the dark color and
bitter flavor associated with products produced from pheno-
lics during processing. Determination of the reactions that
would occur involving these phenolics during processing is
the key to further research on this problem.
In contrast to the effect of pH on the color properties of
oilseed phenolics (15–17), which generally occurs during
a brown solid consisting of at least three undetermined colored protein isolation (18), the effects of heat and pressure are
substances and left a yellow liquid. A newly developed yellow closely related to the steps during oilseed extraction, such as
substance, syringaldehyde, was identified in the liquid phase preheating, cooking, pressing, and desolventization (2,3,13).
by comparing the NMR and mass spectra of the unknown with
those of authentic syringaldehyde. Thomasidioic acid was also
found in the liquid phase. Under the same autoclaving condi-
tions, sinapine bisulfate showed no evidence of any structural
or color changes.
During these steps, heat and pressure are normally required.
In a study on sunflower seed phenolics, heating was found to
decrease the content of simple phenolics (19). Heat treat-
ments were also shown to result in a decrease in the content
of sinapine bisulfate and an increase in lignan content in rape-
seed (20). Heat and pressure seemed to have significant ef-
fects on the color and structure of phenolics. However, the
exact nature of these changes has rarely been reported, partly
because of the complexity of the systems. In addition, there
appeared to be no reports on the coloration of canola pheno-
lics during autoclaving.
Paper no. J8901 in JAOCS 76, 433–441 (April 1999).
KEY WORDS: Autoclaving, canola phenolics, chromato-
graphic analysis, color, sinapic acid, sinapine bisulfate, sy-
ringaldehyde.
Canola provides an excellent edible oil source (1) and is one
Sinapic acid and sinapine bisulfate are the focus of this
of the most important oilseeds in the world (2,3). Canola pro- paper since they are the major simple and esterified pheno-
tein has a well-balanced amino acid content and a favorable lics in canola (9,21–24). The present study reports the struc-
protein efficiency ratio (4,5). Therefore, interest in preparing tural changes and coloration of sinapic acid and sinapine
food-grade protein from canola meal has been increasing bisulfate during autoclaving with a focus on the development
(
6–8). Unfortunately, the use of canola protein is limited by of colored substances. The experiments were performed with
the presence of some antinutritional compounds such as glu- pure phenolic solutions to ensure that the information ob-
cosinolates, phytates and phenolics (9). Phenolics have sig- tained reflected the substances of interest.
nificant effects on the visual attributes, flavor characteristics,
nutritional, and even functional properties of oilseed prod-
ucts (9–12). Color and some other effects are believed to be
MATERIALS AND METHODS
developed during oilseed processing (11). Commercial defat- Sources of materials. Sinapic acid (3,5-dimethoxy-4-hydroxy-
ting of canola is currently accomplished almost exclusively cinnamic acid) and syringaldehyde (4-hydroxy-3,5-
by prepress and solvent extraction systems, with flaking and dimethoxybenzaldehyde) were purchased from Aldrich
cooking of the whole seed before pressing in expellers Chemical Co. (Milwaukee, WI). Syringaldehyde is a dull yel-
low solid (Sigma-Aldrich Canada Ltd., material safety data
sheet, 1997). Aldrich silica gel (230–400 mesh, 60 Å) was
*To whom correspondence should be addressed.
E-mail: arntfie@cc.umanitoba.ca
used for column chromatography. Sinapine was isolated from
Copyright © 1999 by AOCS Press
433
JAOCS, Vol. 76, no. 4 (1999)

434
R. CAI ET AL.
Sinapis alba certified seed from Tilney Mustard crop as was autoclaved for 30 min. Autoclaving was conducted with
sinapine bisulfate according to the method outlined by Clan- an AMSCO Eagle 3000 Series Sterilizer (American Sterilizer
dinin (25). Sinapine bisulfate was used in the place of sinap- Company, Horsham, PA). Time, temperature, and pressure
ine for assessing the autoclaving effect. Acetic acid and were controlled automatically. A gravity model was used for
sodium hydroxide used for high-performance liquid chroma- the autoclaving of the samples, where no liquid water was
tography (HPLC) were verified American Chemical Society added during autoclaving. Temperature and pressure were
(
ACS)-grade and purchased from Fisher Scientific Co. (Ne- routinely set to 121°C (250°F) and 0.1 MPa (15 psi). Spec-
pean, Ontario, Canada). Other chemicals used for HPLC were tral analysis of the samples was conducted both before and
HPLC-grade. All other chemicals, unless stated otherwise, after autoclaving to determine the ultraviolet (UV) spectra
were verified ACS-grade and purchased from Fisher Scien- and Hunter L a b values. After autoclaving, the solution was
tific Co.
filtered through a 0.45-µm filter, giving a brown solid phase
Thomasidioic acid was prepared according to Ahmed et and a yellow liquid phase.
al. (26) and Rubino et al. (27) with some modifications.
Analysis of the brown solid. The solid phase was recov-
Sinapic acid (20 mg) was dissolved in 0.4 mL methanol. This ered from the filter paper by washing with ethyl acetate, dry-
solution was added to an aqueous FeCl solution (40 mg/1.6 ing with MgSO , and evaporating under vacuum to remove
3
4
mL). The solution was aerated for a few seconds. The violet- solvent. HPLC analysis and thin-layer chromatography
red precipitate was collected and treated with concentrated (TLC; 100% ethyl acetate) were conducted on the recovered
H SO for 30 s. After diluting with an equal amount of water, solid samples. Three fractions from the TLC at R 0.95, 0.91,
2
4
f
the resulting mixture was extracted with ethyl acetate (3 and 0.72 were recovered from the silica gel by extracting with
times–5 mL each), dried with MgSO , and evaporated under methanol, filtering through a 0.45-µm filter to remove silica
4
vacuum. The residue was dissolved in methanol, activated gel, and evaporating under vacuum to remove the solvent. A
carbon was added, and the mixture was then filtered through nuclear magnetic resonance (NMR) spectrum analysis was
a 0.45-µm filter. The methanol was evaporated under vac- performed for each of these fractions.
uum. The resulting thomasidioic acid was used as a standard
for the identification of new compounds formed during auto- from the filtration was analyzed immediately following the
claving. Thomasidioic acid is a colorless solid (26). autoclaving and filtration using HPLC analysis. A 10-mL
Analysis of the yellow liquid. The yellow liquid phase
Procedure for determination of structural changes and sample was kept as a control for later HPLC analysis. The
coloration of sinapic acid. The procedure for the determina- substances in the rest of the liquid phase were recovered by
tion of color and structural changes of sinapic acid is shown extracting the solution with ethyl acetate. The organic phases
in Scheme 1. A 100 µg/mL solution was prepared by dissolv- were combined, dried with MgSO , and evaporated under
4
1
ing 200 mg of sinapic acid in 2000 mL distilled water in a vacuum to remove the solvent. After recording its H NMR
3
000-mL beaker. About 10 mL of this solution was kept as a spectrum, the residue, a red solid, was recrystallized from
control sample. In determining the effect of autoclaving time CH Cl . The crystals, which were mainly sinapic acid (TLC,
2
2
on the color properties of the solution, three samples of 10 NMR), were discarded, while the liquid CH Cl solution was
2
2
mL each were taken from the above solution and autoclaved chromatographed through a 300 × 2 cm silica gel column
for 15, 30, and 45 min, respectively. The rest of the solution using ethyl acetate as eluant. The elution was run under slight
air pressure and 20 fractions were collected. These fractions
were evaporated to dryness under a stream of warm air. The
fourth fraction, a dull orange solid, was the major fraction.
This solid was analyzed using NMR spectroscopy, mass spec-
troscopy, and HPLC.
Preparation and treatment of sinapine bisulfate solution.
A 100 µg/mL sinapine bisulfate solution was prepared by dis-
solving 5 mg of sinapine bisulfate in 50 mL distilled water.
While 10 mL of this solution was kept as control, the rest was
autoclaved for 45 min at the same conditions, as for sinapic
acid. Spectral and HPLC analysis were performed for the
control and the autoclaved sample.
Spectral analysis and color determination. Spectral analy-
ses were performed with a Hewlett-Packard 8452 diode array
spectrophotometer (Hewlett-Packard, Palo Alto, CA) with
MS-DOS UV-visible (VIS) software looking at both UV
(200–350 nm) and VIS spectra (380–780 nm). The same sam-
ples were used for both spectral and HPLC analysis. UV
spectra were recorded as curves of absorbance as a function
of wavelength. VIS spectra were recorded as percentage of
SCHEME 1
JAOCS, Vol. 76, no. 4 (1999)

STRUCTURAL CHANGES AND COLORATION OF CANOLA PHENOLICS
435
light transmittance as a function of wavelength. The color of nents, TLC was performed using silica gel TLC plates (Wa-
a solution was determined by taking the transmittance data at terman, Clifton, NJ), with 100% ethyl acetate as elution sol-
the wavelength range from 380 to 780 nm, calculating the vent. Crude reproductions of the results were made using a
Commission Internationale de l’Eclairage (CIE) 1931 tristim- computerized drawing program (Corel Flow, version 7, Corel
ulus values, X, Y, and Z, and mathematically converting to Corporation Limited, Canada). R was measured manually.
f
Hunter L a and b values (28).
Nuclear magnetic resonance and mass spectra. NMR and
The calculation of the CIE 1931 tristimulus values, X, Y, mass spectral analyses were performed on several materials
and Z from the above transmittance data was based on a obtained during the isolation procedure. Samples were dis-
weight-ordinate method with a wavelength interval ∆λ = 10 solved in acetone D , filtered through a pipet filled with
6
(
28). This calculation was done with a computer spreadsheet Kimwipes paper (Kimberly-Clark Corp., Roswell, GA) and
program (Quattro Pro, Version 7, Corel Corporation Limited, collected in NMR tubes for analysis.
The H and 13C NMR spectra were recorded using a
1
Canada). The CIE standard illuminant D65 was used as the il-
luminant in the calculation. Since these X, Y, and Z values do Brucker AM-300 spectrometer (Karlsruhe, Germany) with
not have a direct proportion to real color intensities, they tetramethylsilane as internal standard. Acetone D was used
6
were converted to Hunter L a b values using the equations L as solvent for all of the NMR analyses. Mass spectra were ob-
0
.5
0.5
=
0
10·Y , a = 17.5·(1.02·X − Y)/Y , and b = 7.0·(Y − tained on a vacuum generator, model VG 7070E-HF instru-
0
.5
.847·Z)/Y (28). All the reported values were means of two ment (Manchester, England).
determinations.
HPLC analysis. HPLC was used to monitor the structural
changes in sinapic acid and sinapine bisulfate following the
RESULTS AND DISCUSSION
autoclave treatment. Chromatographic equipment consisted Sinapic acid. Spectral and color changes during autoclaving
of two Waters (Milford, MA) pumps (models 501 and 510), (development of yellow appearance). UV and visual spectral
an automated gradient controller model 680 (Waters), a Shi- changes for sinapic acid solutions before and after autoclav-
madzu (Kyoto, Japan) SPD-6A UV spectrophotometric de- ing are shown in Figure 1 A,B. The UV spectrum of the
tector, and a Hewlett-Packard (Avondale, PA) model sinapic acid solution changed from two major peaks (max-
HP3396II integrator connected with a peak 96 computer soft- ima around 230 and 320 nm) to three major peaks (maxima
ware (Hewlett-Packard Company, Avondale, PA) . A reverse- around 210, 225, and 310 nm). There was an increased ab-
phase C18 column (Supelcosyl, 3-µm particle size, 33 × 4.6 sorbance in the region above 380 nm after autoclaving (Fig.
mm i.d.; Supelco, Bellefonte, PA) was used. The elution sol- 1A). It appears that about 40% of the sinapic acid was con-
vent consisted of two buffers. Buffer A was a 0.05 M sodium
acetate buffer prepared by a 1:100 dilution of a stock pH 4.7
acetate buffer. The stock buffer was prepared by adjusting 5
M acetic acid to pH 4.7 with solid sodium hydroxide (29).
Buffer A was filtered through a 0.45 µm filter. Buffer B was
pure methanol. The column was maintained at 37°C and run
at a constant flow rate of 1.4 mL/min.
For sinapic acid solutions before and after autoclaving, the
initial elution solvent was 15% methanol and 85% buffer A.
After a 10-min isocratic flow, a 2-min linear gradient was
used to change the solvent composition to 100% methanol.
This composition was maintained for 2 min, after which an-
other 2-min linear gradient returned the solvent to its original
composition.
For the brown solid separated by filtration through a 0.45-
µm filter, the initial elution solvent was 15% methanol and
8
5% buffer A. This composition was altered to 100%
methanol with a 10-min linear gradient. The solvent compo-
sition was maintained at 100% methanol for 2 min, after
which a 2-min linear gradient returned the solvent to the orig-
inal composition. The brown solid was dissolved in methanol
for the HPLC analysis. The same solvent system was used for
sinapine bisulfate solutions. Sinapine bisulfate was eluted
within the 10-min linear gradient. Mixtures of syringalde-
hyde and sinapic acid were separated by the same linear gra-
dient solvent.
FIG. 1. (A) Effect of autoclaving (121°C, 0.1 MPa, 15 min) on the ultra-
violet spectrum (diluted 10 times), and (B) visual transmittance spec-
TLC. In order to qualitatively analyze the colored compo- trum of a 100 µg/mL sinapic acid solution.
JAOCS, Vol. 76, no. 4 (1999)

436
R. CAI ET AL.
verted to new substances during autoclaving, and that the UV
spectrum still mainly represented sinapic acid.
The effects of autoclaving on the visual transmittance
spectrum of a 100 µg/mL sinapic acid solution were more
dramatic (Fig. 1B). After autoclaving, the solution exhibited
a decreased percentage transmittance (increased absorbance),
especially in the range from 380 to 480 nm. This would be
consistent with the yellow appearance of the solution, since
the maximum absorbance occurred in the blue wavelength re-
gion. A more precise determination of color was done by cal-
culating the CIE 1931 tristimulus values, X, Y, and Z, and
converting to the Hunter L a b values.
The effect of autoclaving time on the Hunter L a b values
of a 100 µg/mL sinapic acid solution is shown in Figure 2.
Autoclaving treatment decreased whiteness (L value), while
significantly increasing yellow (b value) and slightly increas-
ing green color. As a result, the appearance of the solution
turned from colorless to yellow after autoclaving. Increasing
the heat treatment time from 15 to 30 min resulted in an in-
crease in the intensity of yellow. However, whiteness and
green seemed not to be affected by autoclaving time. In addi- FIG. 3. High-performance liquid chromatography (HPLC) chro-
matograms of the solid phase. Samples were applied to HPLC column
as methanol solutions. (A) control sinapic acid, (B) brown solid sepa-
rated by filtering through a 0.45 µm filter. 1. Sinapic acid, other peaks
were unidentified.
tion, further increases in time after 30 min seemed to have no
significant effects on the yellow intensity (b value).
Before the autoclaved solution was analyzed by HPLC, it
was filtered through a 0.45-µm filter. A brown solid was re-
tained on the filter. This suggested that some of the colored of peaks were present in the HPLC chromatogram (Fig. 3B),
substances, formed during autoclaving, were water insoluble. suggesting the presence of several new compounds. Determi-
Therefore, this filtration was used as a technique to separate nation of the structures of these compounds required further
the brown solid from the yellow liquid, as indicated in experimentation.
Scheme 1.
Based on the HPLC results indicating that there were sev-
Analysis of the brown solid phase. HPLC chromatograms eral compounds present in the solid phase after filtration (Fig.
of standard sinapic acid and the colored substances in the 3A,B), TLC was used to separate three strong bands, which
brown solid phase are shown in Figure 3 A,B. These sub- were yellow, orange and purple (Fig. 4). These three bands
stances were dissolved in methanol for the HPLC analysis. had retention factors (R ) of 0.95 ± 0.01, 0.91 ± 0.01, and 0.72
f
Figure 3A shows the HPLC chromatogram of the standard ± 0.02, respectively, with the yellow band being the major
sinapic acid and Figure 3B that of the brown solid. A number fraction. In addition, it was noted that the yellow and orange
compounds were UV light-sensitive and turned purple when
exposed to UV light for a few minutes. There were also some
other faintly colored bands found using TLC analysis. The
compounds found in the TLC analysis do not necessarily cor-
respond to the compounds detected by HPLC.
The results of NMR analysis of the yellow, orange, and
purple bands gave poorly resolved spectra with many peaks.
FIG. 2. Effect of autoclaving on the Hunter L a b values of a 100 µg/mL
sinapic acid solution. L, white bars; a, gray bars; b, dark bars.
FIG. 4. Thin-layer chromatography of the solid phase with 100% ethyl
acetate as eluant.
JAOCS, Vol. 76, no. 4 (1999)

STRUCTURAL CHANGES AND COLORATION OF CANOLA PHENOLICS
437
The only conclusion that could be drawn was that the yellow autoclaving (Fig. 5B). HPLC results indicated that about 40%
substance seemed to be a dimer of sinapic acid. There ap- of the sinapic acid was lost during autoclaving. Therefore,
peared to be two sets of signals for the protons on the aro- only part of the sinapic acid was converted to new substances
matic ring and two sets of signals for the methoxyl groups for during autoclaving, and these new substances may be present
the yellow substance, in comparison with only one set of each only in small amounts since almost no peak other than sinapic
of these signals for the sinapic acid. It was also possible that acid stood out. After 20 d at 22°C, with daily loosening of the
the yellow substance may be composed of two major prod- cap of the small bottle for air, the peak of thomasidioic acid
ucts as evidenced by signals in both the aromatic and was more pronounced whereas the unknown peak seemed to
methoxyl region.
disappear. The peak height of the sinapic acid also decreased
Analysis of the yellow liquid phase. The results of the (Fig. 5C). The pH of the solution increased from 4.3 to 5.3
HPLC analysis for the control sinapic acid and the liquid during this period of storage. The increase in thomasidioic
phase of the autoclaved samples after filtration are shown in acid content may be due to the conversion of sinapic acid to
Figure 5A,C. Immediately following autoclaving, a small thomasidioic acid at higher pH values, although the final pH
peak corresponding to thomasidioic acid and another peak value was still in the acid region. Rubino et al. (27) reported
(
unknown) were found in the liquid phase (Fig. 5B). Sinapic the formation of thomasidioic acid when sinapic acid was ex-
acid, however, remained as the predominant component after posed to neutral or alkaline conditions. In detailed studies on
the conversion of sinapic acid to thomasidioic acid, it was
shown that oxygen was necessary for the conversion (27,30).
During the storage test, air was introduced occasionally. The
formation of thomasidioic acid was confirmed by NMR spec-
troscopy. Some thomasidioic acid was formed during the au-
toclave treatment, although the mechanism of this conversion
is unknown.
When the sinapic derivatives in the liquid phase were re-
covered and chromatographed on silica gel, a new substance,
1
13
syringaldehyde, was identified by H and C NMR, by mass
spectral analysis, and by HPLC analysis of one of the major
fractions.
The identification of syringaldehyde in the liquid phase
was made by NMR and mass spectroscopic investigation of
the major fraction recovered from silica gel chromatography.
NMR spectral data for standard sinapic acid, NMR and mass
spectral data for standard syringaldehyde, and the spectral
data for the unknown compound present in the liquid are
shown in Table 1. The NMR spectrum and mass spectral data
of the unknown were identical to those of the standard sy-
ringaldehyde. This indicated that syringaldehyde had formed
during autoclaving.
The identification of syringaldehyde was not readily ap-
parent, since the unknown sample was a mixture of sy-
ringaldehyde and several other substances. In fact, the results
1
of H NMR spectrum of the whole residue showed the sam-
ple to be the mixture of three major substances, sinapic acid,
syringaldehyde, and thomasidioic acid. After the residue was
recrystallized from CH Cl and chromatographed through the
2
2
silica gel, the major fraction still contained two substances,
sinapic acid and syringaldehyde (NMR). This meant that the
separation of syringaldehyde from sinapic acid was not com-
plete, although the separation involved several steps. This
may be, in part, due to the similarity of the polar nature for
both substances. Because of the incomplete separation, the
1
FIG. 5. HPLC chromatograms of the yellow liquid phase. Samples were
applied to the HPLC column as an aqueous solution. (A) Control sinapic
acid (sample pH 4.3), (B) liquid phase separated by filtering through a
presence of syringaldehyde was not easily confirmed by H
13
NMR spectroscopy alone. However, C NMR spectroscopy
provided the typical nine-carbon spectrum. In addition, the
0
.45 µm filter (sample pH 4.3), (C) sample of filtered liquid phase after
1
H NMR spectrum exhibited a typical low-field aldehyde
standing at 22°C for 20 d (sample pH 5.3). 1. Sinapic acid, 2. thomasid-
ioic acid, other peaks were unidentified. For abbreviation see Figure 3.
proton signal. With this information at hand, NMR spectra
JAOCS, Vol. 76, no. 4 (1999)

438
R. CAI ET AL.
TABLE 1
Nuclear Magnetic Resonance and Mass Spectra of Standards
and Compounds Produced During Autoclaving
Description
Sinapic acid
Spectrum technique
1H NMR (acetone D6)
Spectra
7.58 (d, 1H, J = 15.9 Hz), 7.01 (s, 2H),
6.39 (d, 1H, J = 15.8 Hz), 3.91 (s, 6H,
OCH3)
13C NMR (acetone D6)
167.98 (CO), 148.39 (2C),146.00 (C),
1
1
39.20 (C), 126.01 (CH),116.08 (CH),
06.63 (2CH), 56.52 (2CH3)
Syringaldehyde
1H NMR (acetone D6)
13C NMR (acetone D6)
9.79 (s, 1H), 7.21 (s, 2H),
.90 (s, 6H, OCH3)
191.12 (CO), 148.94 (2C),142.93 (C),
3
1
5
28.96 (C),107.73 (2CH),
6.63 (2CH3)
Mass spectrum m/z
182 (100), 167 (20), 153 (8), 139 (13),
(relative intensity)
111 (16), 96 (11), 79 (12), 65 (15)
Compound identified
1H NMR (acetone D6)
13C NMR (acetone D6)
9.81 (s, 1H), 7.23 (s, 2H),
3.92 (s, 6H, OCH3)
190.93 (CO), 148.84 (2C), 142.82 (C),
1
5
28.87 (C), 107.63 (2CH),
6.73 (2CH3)
Mass spectrum m/z
182 (100), 167 (20), 153 (7), 139 (13),
(relative intensity)
111 (20), 96 (9), 79 (15), 65 (18)
were examined more closely. Two groups of peaks, represent- 6C). A sinapic acid standard (not shown, result similar to Fig.
ing sinapic acid and syringaldehyde with a molar ratio of 6A) and a syringaldehyde standard (Fig. 6B) were also ex-
1
13
around 1:0.27, were found in both H and C NMR spectra amined separately to ensure that no interaction of the two
for the major fraction from silica gel chromatography. Three substances was interfering with the elution times. To prove
groups of peaks, representing sinapic acid, syringaldehyde that syringaldehyde was produced during autoclaving but not
molar and thomasidioic acid with a ratio of around 1:0.1:0.03, during the sample preparation following the autoclaving, the
1
were found in H NMR spectrum of the ethyl acetate extract standard sinapic acid was used as a blank sample and run
from the liquid phase. The major fraction from silica gel through the same experimental procedure but without auto-
chromatography was also analyzed using mass spectroscopy. claving. To do this, the sinapic acid was dissolved in water,
The major mass peaks corresponded to syringaldehyde alone then recovered by ethyl acetate extraction followed by recrys-
since syringaldehyde has a lower boiling point than does tallization from CH Cl . After removing the solvent under
2
2
sinapic acid. Only syringaldehyde was evaporated and vacuum, the substance was analyzed by HPLC. The HPLC
recorded in the spectrum under the test condition. The spec- chromatogram of this blank sinapic acid is given in Figure
trum of the unknown was almost perfectly matched with that 6A. The single peak was identified as sinapic acid. There was
of the standard. The unknown was therefore identified to be no evidence of syringaldehyde in the control sample. This
syringaldehyde.
proved that syringaldehyde was formed during autoclaving.
Based on NMR and mass spectral results, HPLC analysis Attempts were made to improve resolution of the HPLC
was conducted to confirm the presence of syringaldehyde in analysis by using two other gradients for the mixture of
the liquid phase. To ensure that the formation of syringalde- sinapic acid and syringaldehyde. However, similar results
hyde was not due to other treatments following the autoclav- with no improvement in resolution were obtained. Further ex-
ing, a blank control sinapic acid was also analyzed. Figure perimentation for this purpose was terminated since the pres-
6A–D shows the results of HPLC analysis for a recrystallized ence of syringaldehyde was firmly confirmed based on NMR
control sinapic acid (Fig. 6A), a standard sample of sy- and mass spectral analysis. HPLC analysis was mainly con-
ringaldehyde, (Fig. 6B), the mixture of the standard sinapic ducted to confirm the results of the NMR and mass spectral
acid, and the standard syringaldehyde (Fig. 6C), and the un- analysis. Thomasidioic acid found in the previous HPLC
known sample from the column chromatographic step (Fig. chromatograms (Fig. 5A–C) was not detected in the major
6
D). Two peaks were found in the unknown sample, which fraction from silica gel column chromatography. This may
were identified as sinapic acid and syringaldehyde. The iden- indicate a good separation between thomasidioic acid and sy-
tification was done by comparing the chromatogram of the ringaldehyde during the silica gel chromatography.
unknown sample (Fig. 6D) with that of the mixture of the
Retention time for syringaldehyde was very close to that
standard sinapic acid and the standard syringaldehyde (Fig. for sinapic acid. In addition, the concentration of syringalde-
JAOCS, Vol. 76, no. 4 (1999)

STRUCTURAL CHANGES AND COLORATION OF CANOLA PHENOLICS
439
hyde in the liquid phase was low. It was estimated from about 6 µg/mL syringaldehyde, about 2 µg/mL thomasidioic
HPLC data that 40% of the sinapic acid was converted to acid (HPLC, NMR), and the rest were undetermined sub-
other substances during a 30-min autoclaving. The starting stances. The earlier HPLC analysis of the yellow liquid phase
material was a 100 µg/mL sinapic acid aqueous solution. The (Fig. 5B,C) did not show the presence of syringaldehyde.
autoclaved solution contained about 60 µg/mL sinapic acid, This is due to the excess of sinapic acid present, which masks
the signal from syringaldehyde owing to their similar reten-
tion time.
In order to evaluate the color significance of syringalde-
hyde, the UV and transmittance spectra of syringaldehyde
were compared with those of the sinapic acid (Fig.7 A,B).
The UV spectra for both substances are shown in Figure 7A.
Syringaldehyde had a spectrum with two maxima of around
220 and 310 nm. Both maxima are at lower wavelengths than
those of the sinapic acid. While sinapic acid shows almost no
absorbance in the wavelength region from 275 to 400 nm, sy-
ringaldehyde shows a slight absorbance in this region. This is
consistent with the dull yellow appearance of syringaldehyde.
The transmittance spectra for syringaldehyde and sinapic acid
are shown in Figure 7B. Syringaldehyde shows a relatively
lower percentage transmittance than does sinapic acid in the
whole visual light region, especially in the wavelength region
from 380 to 400 nm. Therefore, the formation of syringalde-
hyde would contribute a yellowness and a darkness to the
sinapic acid solution. However, comparison of the transmit-
tance spectrum of syringaldehyde solution (Fig. 7B) with that
of the autoclaved sinapic acid solution (Fig. 1B), leads to the
FIG. 6. HPLC chromatograms of the major fraction from silica gel chro-
matography (A) sinapic acid blank, (B) standard syringaldehyde, (C)
mixture of standard sinapic acid and syringaldehyde, and (D) unknown FIG. 7. (A) Comparison of ultraviolet spectra (dilute 10 times and (B) vi-
sample identified to be a mixture of sinapic acid and syringaldehyde. 1. sual transmittance spectra of 100 µg/mL syringaldehyde and 100 µg/mL
Sinapic acid. 2. Syringaldehyde. For abbreviation see Figure 3.
sinapic acid aqueous solutions.
JAOCS, Vol. 76, no. 4 (1999)

440
R. CAI ET AL.
conclusion that syringaldehyde contributes only a very small
amount of yellowness and darkness to the solution. Other un-
determined components, especially those three colored sub-
stances shown by TLC, could be the major components re-
sponsible for color change of sinapic acid during autoclav-
ing. The concentration of these highly colored compounds
may be much lower than that of syringaldehyde.
Sinapine bisulfate. The results of UV and visual spectral
analysis for sinapine bisulfate solution before and after auto-
claving are shown in Figure 8 A,B. Spectra of sinapine bisul-
fate for the control and autoclaved sample were perfectly
matched. This would suggest that no changes occurred dur-
ing autoclaving.
Both control and autoclaved sinapine bisulfate solutions
(100 µg/mL) were analyzed using HPLC for a storage period
of 20 d at 22°C with a daily loosening of the cap of the small
bottle for air. No differences were found between the two
groups of samples (HPLC chromatograms not shown). This
also suggested that no structural changes occurred during au-
toclaving.
In conclusion, it has been demonstrated that a yellow col-
oration developed when a 100 µg/mL sinapic acid aqueous so-
SCHEME 2
lution was autoclaved for 15 min at 121°C and 0.1 MPa. A sent by TLC analysis. However, their identities have not been
newly formed yellow substance was identified to be sy- determined. The structural and color changes of a 100 µg/mL
ringaldehyde by NMR, mass spectral, and HPLC analysis. sinapic acid aqueous solution during autoclaving are summa-
Thomasidioic acid was also formed during autoclaving. There rized in Scheme 2. Sinapine bisulfate, under the same condi-
were at least three other colored substances estimated to be pre- tions, showed no evidences of any structural or color changes.
ACKNOWLEDGMENT
Financial support of this research by the Natural Sciences and Engi-
neering Research Council of Canada and the award of the J.W. Bar-
low Graduate Fellowship to Mr. R. Cai are gratefully acknowledged.
REFERENCES
1
. Ackman, R.G., Canola Fatty Acids—An Ideal Mixture for
Health, Nutrition, and Food Use, in Canola and Rapeseed: Pro-
duction, Chemistry, Nutrition and Processing Technology,
edited by F. Shahidi, Van Nostrand Reinhold, New York, 1990,
pp. 81–98.
2
. Eskin, N.A.M., B.E. McDonald, R. Przybylski, L.J. Malcolm-
son, R. Scarth, T. Mag, K. Ward, and D. Adolph, Canola Oil.
Offprints from Bailey’s Industrial Oil and Fat Products, Vol. 2,
edited by Y.H. Hui, John Wiley & Sons, Inc., New York, 1996,
pp. 1–95.
3
4
. Shahidi, F., Canola and Rapeseed: Production, Chemistry, Nu-
trition and Processing Technology, Van Nostrand Reinhold,
New York, 1990, pp. 3–13.
. Liu, R.F.K., L.U. Thompson, and J.D. Jones, Yield and Nutri-
tive Value of Rapeseed Protein Concentrate, J. Food Sci.
4
7:977–981 (1982).
5
6
. Ohlson, R., and K. Anjou, Rapeseed Protein Products, J. Am.
Oil Chem. Soc. 56:431–437 (1976).
. Ismond, M.A.H., and W.D. Welsh, Application of New Method-
ology to Canola Protein Isolation, Food Chem. 45:125–127
(1992).
7
. Gillberg, L., and B. Tornell, Preparation of Rapeseed Protein
Isolate. Dissolution and Precipitation Behavior of Rapeseed Pro-
teins, J. Food Sci. 41:1063–1069 (1976).
FIG. 8. (A) Effect of autoclaving on the UV spectrum (diluted 10 times,
and (B) visual transmittance spectrum of a 100 µg/mL sinapine bisulfate
bisulfate aqueous solution.
8. Owen, D.F., C.O. Chichester, J. Granadino, and F. Monckeberg,
JAOCS, Vol. 76, no. 4 (1999)

STRUCTURAL CHANGES AND COLORATION OF CANOLA PHENOLICS
441
A Process for Producing Nontoxic Rapeseed Protein Isolate and 20. Jensen, S.K, H.S. Olsen, And H. Sorensen, Aqueous Enzymatic
an Acceptable Feed By-Product, Cereal Chem. 48:91–96
1971).
. Shahidi, F., and M. Naczk, An Overview of Phenolics of Canola
and Rapeseed: Chemical, Sensory and Nutritional Significance,
J. Am. Oil Chem. Soc. 69:917–924 (1992).
0. Shahidi, F., and M. Naczk, Food Phenolics: Sources, Chemistry,
Effects and Applications, Technomic Publishing Company, Inc.,
Lancaster, Pa., 1995.
Processing of Rapeseed for Production of High Quality Prod-
ucts, in Canola and Rapeseed: Production, Chemistry, Nutri-
tion and Processing Technology, edited by F. Shahidi, Van Nos-
trand Reinhold, New York, 1990, pp. 331–343.
(
9
21. Blair, R., and R.D. Reichert, Carbohydrate and Phenolic Con-
stituents in a Comprehensive Range of Rapeseed and Canola
Fractions: Nutritional Significance for Animals, J. Sci. Food
Agric. 35:29–35 (1984).
1
1
1. Youngs, C.G., Technical Status Assessment of Food Protein 22. Sosulski, W.F., and K.J. Dabrowski, Composition of Free and
from Canola, in Research on Canola Seed, Oil and Meal, 9th
Project Report, Canola Council of Canada, Winnipeg, Canada,
Hydrolyzable Phenolic Acids in the Flours and Hulls of Ten
Legume Species, J. Agric. Food Chem. 32:131–133 (1984).
23. Krygier, K., F. Sosulski, and L. Hogge, Free, Esterified and In-
soluble Phenolic Acids. II. Composition of Phenolic Acids in
Rapeseed Flour and Hulls, Ibid 30:334–336 (1982).
24. Fenton, T.W., T. Leung, and D.R. Clandinin, Phenolic Compo-
nents of Rapeseed Meal, J. Food Sci. 45:1702–1705 (1980).
25. Clandinin, D.R., Rapeseed Oil Meal Studied. 4. Effect of Sinap-
ine, the Bitter Substance in Rapeseed Oil Meal, on the Growth
of Chickens, Poult. Sci. J. 40:484–487 (1961).
26. Ahmed, R., M. Leherer, and R. Stevenson, Synthesis of Thomasic
Acid, Tetrahedron 29:3753–3759 (1973).
27. Rubino, M.I., S.D. Arntfield, and J.L. Charlton, Conversion of
Phenolics to Lignans: Sinapic Acid to Thomasidioic Acid, J.
Am. Oil Chem. Soc. 72:1465–1470 (1995).
1
991, pp. 309–341.
1
2. Rubino, M.I., S.D. Arntfield, C.A. Nadon, and A. Bernatsky,
Phenolic Protein Interactions in Relation to the Gelation Proper-
ties of Canola Protein, Food Res. Int. 29:653–659 (1996).
3. Hamilton, R.J., and A. Bahail, Fats and Oils: Chemistry and
Technology, Elsevier Applied Science, London, 1987.
4. Buhr, N., Basic Principles and Modern Practices, in Edible Fats
and Oils Processing, edited by D.R. Erickson, American Oil
Chemists’ Society, Champaign, 1989, pp. 43–48.
5. Ribereau-Gayon, P., Plant Phenolics, Oliver & Boyd, Edin-
burgh, 1972, pp. 81–105.
6. Harborne, J.B., Ultraviolet Spectroscopy of Polyphenols, in
Methods in Polyphenolics, edited by J.B. Pridham, The Macmil-
lan Company, New York, 1964, pp. 13–36.
1
1
1
1
28. Judd, D. B., and G. Wyzecki, Color in Business, Science and In-
dustry, 3rd edn., John Wiley & Sons, New York, 1975, pp. 144,
318.
1
7. Austin, F.L., and I.A. Wolff, Sinapine and Related Esters in
Seed Meal of Crambe Abyssinica, J. Agric. Food Chem.
1
6:132–135 (1968).
29. Hagerman, A.E., and R.L. Nicholson, High Performance Liquid
Chromatographic Determination of Hydroxycinnamic Acids in
the Maize Mesocotyl, J. Agric. Food Chem. 30:1098–1102
(1982).
30. Lee, K.S. Oxidative Coupling of Sinapic Acid. MSc. Thesis,
University of Manitoba, Winnipeg, Manitoba, Canada, 1997.
1
8. Keshavarz, E., R.K.M. Cheung, R.C.M. Liu, and S. Nakai,
Adaptation of the Three Stage Extraction Process to Rapeseed
Meal for Preparation of Colorless Protein Extracts, Can. Inst.
Food Sci. Technol. J. 10:73–77 (1977).
9. Sastry, M.C.S., and N. Subramanian, Effect of Heat Processing
on Phenolic Constituents and Nutritional Quality of Sunflower
Flours, J. Am. Oil Chem. Soc. 62:1131–1134 (1985).
1
[Received June 12, 1998: accepted January 9, 1998]
JAOCS, Vol. 76, no. 4 (1999)