Stability of capsinoid in various solvents

Article abstract of DOI:10.1021/jf0103424

To investigate the stability of capsinoid in solvents, the quantitative change of vanillyl nonanoate, a synthetic model capsinoid, in various solvents was measured by HPLC. Vanillyl nonanoate was stable in nonpolar solvents, whereas it was labile in polar solvents. In particular, vanillyl nonanoate tended to decompose in protic solvents such as alcohol and water. Structures of the decomposition products from vanillyl nonanoate in methanol and ethanol were determined to be methyl and ethyl vanillyl ethers, respectively. To clarify the decomposition mechanism of capsinoid, six analogues of vanillyl nonanoate were tested. The stability of the analogues in organic solvents suggested that the hydroxyl group in the para-position of the benzene ring largely contributes to the decomposition of capsinoid.

Full text of DOI:10.1021/jf0103424

4026
J. Agric. Food Chem. 2001, 49, 4026−4030
Sta bility of Ca p sin oid in Va r iou s Solven ts
Kouzou Sutoh, Kenji Kobata, and Tatsuo Watanabe*
Graduate School of Nutritional and Environmental Sciences, University of Shizuoka, 52-1 Yada,
Shizuoka 422-8526, J apan
To investigate the stability of capsinoid in solvents, the quantitative change of vanillyl nonanoate,
a synthetic model capsinoid, in various solvents was measured by HPLC. Vanillyl nonanoate was
stable in nonpolar solvents, whereas it was labile in polar solvents. In particular, vanillyl nonanoate
tended to decompose in protic solvents such as alcohol and water. Structures of the decomposition
products from vanillyl nonanoate in methanol and ethanol were determined to be methyl and ethyl
vanillyl ethers, respectively. To clarify the decomposition mechanism of capsinoid, six analogues of
vanillyl nonanoate were tested. The stability of the analogues in organic solvents suggested that
the hydroxyl group in the para-position of the benzene ring largely contributes to the decomposition
of capsinoid.
Keyw or d s: Capsinoid; capsiate; vanillyl nonanoate; stability; solvolysis
INTRODUCTION
MATERIALS AND METHODS
Ch em ica ls. Capsaicin, vanillyl alcohol, nonanoyl chloride,
3-methoxybenzyl alcohol, o-hydroxybenzyl alcohol, and m-
hydroxybenzyl alcohol were purchased from Sigma-Aldrich
J apan K.K. (Tokyo, J apan). Acetic anhydride, 3,4-dimethoxy-
benzyl alcohol, and p-hydroxybenzyl alcohol were obtained
from Wako Pure Chemical Industries (Osaka, J apan). Capsiate
was isolated from the fruits of C. annuum L. cv. CH-19 Sweet
(1). The rest of the chemicals used in this study were of
analytical grade.
In str u m en ts. Spectroscopic measurements were done with
the following instruments: NMR, Alpha-400 (J EOL, J apan),
1H 399.65 MHz, 13C 100.40 MHz, CDCl3, TMS as internal
standard; HRMS, J MS-AX500 (J EOL); HPLC system com-
posed of a pump, PU-980 (J asco, J apan), a fluorescence
detector, FP-920 (J asco), and a UV detector, UV-970 (J asco).
Sta bility of Ca p sin oid in Solven ts. A methanol solution
(1 mL) of 0.5 mM capsaicin was added as an internal standard
into a 5-mL volume screw-top vial, and the methanol was
evaporated under a nitrogen stream. Then, capsinoid (0.8 mg)
was added to the vial and a 2 mM capsinoid solution was
prepared with addition of the testing solvents. Physiological
saline solution containing 2% ethanol and 10% Tween 80 was
used as an aqueous solvent (aqueous surfactant vehicle).
Immediately after the preparation, the samples were kept at
25 °C. An aliquot of the sample was taken at adequate time
intervals and was subjected to HPLC analysis.
HPLC was carried out under the following conditions:
column, J ’sphere ODS-H80, 150 mm × 4.6 mm i.d. (YMC);
solvent, 80% methanol; flow rate, 0.5 mL/min; detection,
fluorescence, excitation 280 nm, emission 320 nm; UV, 280
nm.
Recently three homologues of nonpungent compounds,
capsiate, dihydrocapsiate, and nordihydrocapsiate, have
been isolated from the fruits of a sweet cultivar of
pepper, CH-19 Sweet (Capsicum annuum L.) (1, 2). They
were fatty acid esters of vanillyl alcohol, and the
compound group was named capsinoid (2). The major
natural capsinoid was capsiate [4-hydroxy-3-methoxy-
benzyl (E)-8-methyl-6-nonenoate (1)] (Figure 1). Cap-
saicinoid is a pungent principal of Capsicum plants (3,
4), and it has various physiological activities (5, 6).
Capsinoid bears a marked structural resemblance to
capsaicinoid, except for the central linkage; that is, an
amide moiety is found in capsaicinoid, and an ester
moiety is found in capsinoid. Capsinoid has no pungency
upon oral tasting. Therefore, it is expected that capsi-
noid has physiological functions similar to those of
capsaicinoid and causes less harmful stimulation. Such
characteristics are important for its wide usage as a food
or medicine. It has been reported that in a feeding
experiment of dry fruits of CH-19 Sweet capsinoid raises
the body temperature in humans (7). However, as
compared with capsaicinoid, capsinoid easily decom-
poses in methanol or during the purification process by
silica gel column chromatography. The decomposition
mechanism is not understood. To undertake further
experimentation in which capsinoid will be used as a
solution, it is necessary to know its stability in various
solvents.
To investigate the stability of capsinoid in various
solvents, the quantitative change of vanillyl nonanoate
(2), a synthetic model capsinoid, in various solvents was
measured by HPLC. The decomposition mechanism of
capsinoid was determined from the structures of the
decomposition products from vanillyl nonanoate and the
stability of synthesized vanillyl nonanoate analogues
having varied substituents in the aromatic portion.
Ch em ica l Syn th eses of Va n illyl Non a n oa te An a logu es.
Vanillyl Nonanoate (2). Vanillyl alcohol (1543 mg) was dis-
solved in dehydrated pyridine (10 mL), and nonanoyl chloride
(0.9 mL) was added dropwise to the solution under ice cooling.
The mixture was stirred for 2 h under ice cooling, followed by
the addition of H2O (20 mL) to terminate the reaction and 2
N HCl (20 mL) to acidify the solution. The solution was
extracted three times with 60 mL each of ethyl acetate, and
the resulting organic portion was washed with water and
dehydrated with sodium sulfate. After evaporation, the residue
was rapidly fractionated by silica gel column chromatography
under the following conditions: silica gel 60 (Merck), 30 g;
column, 800 mm × 36 mm i.d.; eluent, hexane/ethyl acetate
* Author to whom correspondence should be addressed (tele-
phone +81-54-264-5543; fax +81-54-264-5550; e-mail watanbt@
u-shizuoka-ken.ac.jp).
10.1021/jf0103424 CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/26/2001

Stability of Capsinoid in Various Solvents
J. Agric. Food Chem., Vol. 49, No. 8, 2001 4027
F igu r e 1. Structures of capsiate (1), vanillyl nonanoate (2), 4-acetoxy-3-methoxybenzyl nonanoate (3), 3,4-dimethoxybenzyl
nonanoate (4), 3-methoxybenzyl nonanoate (5), o-hydroxybenzyl nonanoate (6), m-hydroxybenzyl nonanoate (7), p-hydroxybenzyl
nonanoate (8), and decomposition products methyl vanillyl ether (9) and ethyl vanillyl ether (10).
(8:2). The fraction containing the desired compound was
purified by reversed-phase column chromatography under the
following conditions: Wakosil 25C18 (Wako), 20 g; column, 120
mm × 21 mm i.d.; eluent, methanol/H2O (4:1). The eluent gave
a colorless oil (2, 191.6 mg, 13% yield). Spectral data of HRMS
and 1H NMR of 2 completely agreed with published data (8).
4-Acetoxy-3-methoxybenzyl Nonanoate (3). The mixture of
vanillyl nonanoate (9.68 mg), acetic anhydride (10 mL), and
dehydrated pyridine (10 mL) was heated for 20 min under
reflux, and then the solvent was evaporated in vacuo. The
residue was purified by HPLC under the following condi-
tions: column, J ’sphere ODS-H80, 150 mm × 20 mm i.d.
(YMC); solvent, 80% methanol; flow rate, 8.0 mL/min; detec-
tion, UV 280 nm. The fraction gave a colorless oil (3, 9.63 mg,
87% yield). Spectral data of 3 were as follows: HRMS, m/z
(M+) calcd for C19H28O5 336.1937, found 336.1956; 1H NMR δ
7.01 (1H, d, J ) 7.6 Hz, 5′-H), 6.94 (1H, dd, J ) 7.6, 1.6 Hz,
6′-H), 6.93 (1H, d, J ) 1.6 Hz, 2′-H), 5.08 (2H, s, 1′-CH2), 3.84
(3H, s, 3′-OCH3), 2.35 (2H, t, J ) 7.6 Hz, 2-H), 2.31 (3H, s,
4′-OCOCH3), 1.64 (2H, quint, J ) 7.6 Hz, 3-H), 1.28 (10H, m,
4-8-H), 0.88 (3H, t, J ) 6.8 Hz, 9-H); 13C NMR δ 173.6 (C-1),
169.0 (4′-OCOCH3), 151.1 (C-3′), 139.6 (C-4′), 135.1 (C-1′), 122.8
(C-5′), 120.6 (C-6′), 112.4 (C-2′), 65.7 (1′-CH2), 55.9 (3′-OCH3),
34.3 (C-2), 31.8 (C-7), 29.2 (C-4), 29.1 (C-5,6), 25.0 (C-3), 22.6
(C-8), 20.7 (4′-OCOCH3), 14.1 (C-9).
similar to that used for 4. Spectral data of 5 were as follows:
HRMS, m/z (M+) calcd for C17H26O3 278.1882, found 278.1870;
1H NMR δ 7.27 (1H, t, J ) 7.8 Hz, 5′-H), 6.90 (3H, m, 2′,4′,6′-
H), 5.09 (2H, s, 1′-CH2), 3.81 (3H, s, 3′-OCH3), 2.36 (2H, t, J )
7.6 Hz, 2-H), 1.64 (2H, quint, J ) 7.6 Hz, 3-H), 1.26 (10H, m,
4-8-H), 0.87 (3H, t, J ) 6.8 Hz, 9-H); 13C NMR δ 173.7 (C-1),
159.7 (C-3′), 137.7 (C-1′), 129.6 (C-5′), 120.3 (C-6′), 113.6 (C-
2′,4′), 65.9 (1′-CH2), 55.2 (3′-OCH3), 34.4 (C-2), 31.8 (C-7), 29.2
(C-4), 29.1 (C-5,6), 25.0 (C-3), 22.6 (C-8), 14.1 (C-9).
o-Hydroxybenzyl Nonanoate (6). o-Hydroxybenzyl nonanoate
(6, 985.3 mg, yield 35%) was obtained from o-hydroxybenzyl
alcohol (2 g) and nonanoyl chloride (1.997 mL) under ice
cooling in a manner similar to that used for 4. Spectral data
of 6 were as follows: HRMS, m/z (M+) calcd for C16H24O3
1
264.1725, found 264.1720; H NMR δ 7.88 (1H, s, OH), 7.27
(2H, m, 4′,6′-H), 6.92 (2H, m, 3′,5′-H), 5.12 (2H, s, 1′-CH2), 2.35
(2H, t, J ) 7.6 Hz, 2-H), 1.61 (2H, quint, J ) 7.6 Hz, 3-H),
1.25 (10H, m, 4-8-H), 0.87 (3H, t, J ) 7.0 Hz, 9-H); 13C NMR
δ 176.6 (C-1), 155.6 (C-2′), 132.2 (C-6′), 131.1 (C-4′), 121.8 (C-
1′), 120.5 (C-5′), 117.9 (C-3′), 63.2 (1′-CH2), 34.2 (C-2), 31.8 (C-
7), 29.1 (C-4), 29.0 (C-5,6), 24.8 (C-3), 22.7 (C-8), 14.1 (C-9).
m-Hydroxybenzyl Nonanoate (7). m-Hydroxybenzyl nonanoate
(7, 1252.5 mg, 44% yield) was obtained from m-hydroxybenzyl
alcohol (2 g) and nonanoyl chloride (2.017 mL) under ice
cooling in a manner similar to that used for 4. Spectral data
of 7 were as follows: HRMS, m/z (M+) calcd for C16H24O3
264.1726, found 264.1707; 1H NMR δ 7.22 (1H, t, J ) 7.6 Hz,
5′-H), 6.90 (1H, d, J ) 7.6 Hz, 6′-H), 6.81 (2H, m, 2′,4′-H), 5.07
(2H, s, 1′-CH2), 2.36 (2H, t, J ) 7.6 Hz, 2-H), 1.64 (2H, quint,
J ) 7.6 Hz, 3-H), 1.26 (10H, m, 4-8-H), 0.87 (3H, t, J ) 7.0
Hz, 9-H); 13C NMR δ 174.1 (C-1), 155.9 (C-3′), 137.8 (C-1′),
129.8 (C-5′), 120.3 (C-6′), 115.2 (C-2′), 115.0 (C-4′), 65.9 (1′-
CH2), 34.4 (C-2), 31.8 (C-7), 29.2 (C-4), 29.1 (C-5,6), 25.0 (C-
3), 22.6 (C-8), 14.1 (C-9).
3,4-Dimethoxybenzyl nonanoate (4). 3,4-Dimethoxybenzyl
nonanoate was synthesized from 3,4-dimethoxybenzyl alcohol
(0.8 mL) and nonanoyl chloride (0.895 mL) at room tempera-
ture in a manner similar to that used for 2. The resulting
residue was purified by silica gel column chromatography
under the following conditions: silica gel 60 (Merck), 30 g;
column, 800 mm × 36 mm i.d.; eluent, hexane/ethyl acetate
(91:9). The eluent gave a colorless oil (4, 1342.6 mg, 92% yield).
Spectral data of 4 were as follows: HRMS, m/z (M+) calcd for
1
C
18H28O4 308.1988, found 308.1963; H NMR δ 6.88 (3H, m,
p-Hydroxybenzyl Nonanoate (8). p-Hydroxybenzyl nonanoate
(8, 1792.4 mg, 65% yield) was obtained from p-hydroxybenzyl
alcohol (2 g) and nonanoyl chloride (1.956 mL) under ice
cooling in a manner similar to that used for 4. Spectral data
of 8 were as follows: HRMS, m/z (M+) calcd for C16H24O3
264.1831, found 264.1805; 1H NMR δ 7.23 (2H, d, J ) 8.4 Hz,
3′,5′-H), 6.82 (2H, d, J ) 8.4 Hz, 2′,6′-H), 5.64 (1H, s, OH),
5.04 (2H, s, 1′-CH2), 2.33 (2H, t, J ) 7.6 Hz, 2-H), 1.62 (2H,
quint, J ) 7.6 Hz, 3-H), 1.26 (10H, m, 4-8-H), 0.87 (3H, t,
J ) 6.8 Hz, 9-H); 13C NMR δ 174.2 (C-1), 155.9 (C-4′), 130.2
2′,5′,6′-H), 5.05 (2H, s, 1′-CH2), 3.88 (3H, s, 4′-OCH3), 3.87 (3H,
s, 3′-OCH3), 2.33 (2H, t, J ) 7.6 Hz, 2-H), 1.63 (2H, quint, J )
7.6 Hz, 3-H), 1.26 (10H, m, 4-8-H), 0.87 (3H, t, J ) 7.0 Hz,
9-H); 13C NMR δ 173.8 (C-1), 149.1 (C-4′), 149.0 (C-3′), 128.8
(C-1′), 121.2 (C-6′), 111.8 (C-5′), 111.1 (C-2′), 66.2 (1′-CH2), 56.0
(4′-OCH3), 55.9 (3′-OCH3), 34.4 (C-2), 31.8 (C-7), 29.2 (C-4),
29.1 (C-5,6), 25.0 (C-3), 22.7 (C-8), 14.1 (C-9).
3-Methoxybenzyl Nonanoate (5). 3-Methoxybenzyl nonanoate
(5, 1398.1 mg, 95% yield) was obtained from 3-methoxybenzyl
alcohol (1.0 mL) and nonanoyl chloride (0.987 mL) in a manner

4028 J. Agric. Food Chem., Vol. 49, No. 8, 2001
Sutoh et al.
(C-2′,6′), 128.2 (C-1′), 115.4 (C-3′,5′), 66.0 (1′-CH2), 34.5 (C-2),
31.8 (C-7), 29.2 (C-4), 29.1 (C-5,6), 25.0 (C-3), 22.6 (C-8), 14.1
(C-9).
Decom p osition of Va n illyl Non a n oa te w ith Meth a n ol.
Vanillyl nonanoate (30.05 mg) was dissolved in methanol (30
mL), and the solution was kept at room temperature. After
the complete decomposition of the vanillyl nonanoate had been
confirmed by HPLC analysis, the solvent was evaporated in
vacuo. Then chloroform (10 mL) was added to the residue. The
resulting white insoluble precipitate was collected by centrifu-
gation. The supernatant was purified by activated alumina
column chromatography under the following conditions: ac-
tivated alumina (Wako), 4 g; column, 500 mm × 20 mm i.d.;
eluent, methanol (50 mL). The eluent gave a colorless oil (9,
9.01 mg).
Methyl Vanillyl Ether (9). Spectral data of 9 were as
follows: HRMS, m/z (M+) calcd for C9H12O3 168.0786, found
168.0742; 1H NMR δ 6.85 (3H, m, 2′, 5′, 6′-H), 5.60 (1H, s, OH),
4.37 (2H, s, 1′-CH2), 3.90 (3H, s, OCH3), 3.36 (3H, s, 3′-OCH3);
13C NMR δ 146.6 (C-3′), 145.3 (C-4′), 130.1 (C-1′), 121.2 (C-6′),
114.1 (C-5′), 110.5 (C-2′), 74.8 (1′-CH2), 57.8 (OCH3), 55.9 (3′-
OCH3).
Decom p osition of Va n illyl Non a n oa te w ith Eth a n ol.
Vanillyl nonanoate (28.01 mg) was dissolved in ethanol (30
mL), and the solution was kept at room temperature. The
subsequent procedures were performed in much the same way
as the decomposition of vanillyl nonanoate with methanol. The
eluent gave a pale yellow oil (10, 17.75 mg).
Ethyl Vanillyl Ether (10). Spectral data of 10 were as
follows: HRMS, m/z (M+) calcd for C10H14O3 182.0943, found
1
182.0911; H NMR δ 6.85 (3H, m, 2′,5′,6′-H), 4.42 (2H, s, 1′-
CH2), 3.90 (3H, s, 3′-OCH3), 3.52 (2H, q, J ) 7.2 Hz, OCH2),
1.23 (3H, t, J ) 7.2 Hz, CH3); 13C NMR δ 146.6 (C-3′), 145.2
(C-4′), 130.5 (C-1′), 122.0 (C-6′), 121.1 (C-5′), 110.6 (C-2′), 72.8
(1′-CH2), 65.5 (OCH2), 55.9 (3′-OCH3), 15.2 (CH3).
F igu r e 2. Change in vanillyl nonanoate with time in various
solvents (A) and several solvents containing water (B) at 25
°C: (A) (0) nonsolvent, (O) ethyl acetate, (]) chloroform, (4)
hexane, (3) acetone, ([) dimethyl sulfoxide, (b) methanol, (2)
ethanol, and (9) aqueous surfactant vehicle (0.9% NaCl
aqueous solution containing 2% ethanol and 10% Tween 80);
(B) (0) ethyl acetate with 1% water, (O) water-saturated ethyl
acetate, (b) methanol with 1% water, (2) ethanol with 1%
water, and (9) methanol with 50% water.
RESULTS
Sta bility of Va n illyl Non a n oa te in Va r iou s Sol-
ven ts. Figure 2A shows the change in vanillyl nonanoate
with time in various solvents at 25 °C. Vanillyl nonanoate
was stable in ethyl acetate, chloroform, hexane, acetone,
and nonsolvent conditions over 2 weeks. In contrast,
vanillyl nonanoate was labile in methanol, ethanol,
dimethyl sulfoxide (DMSO), and aqueous surfactant
vehicle, and the half-life periods of vanillyl nonanoate
in these solvents were about 90, 70, 400, and 115 h,
respectively. From these data, vanillyl nonanoate was
stable in nonpolar solvents, whereas it was labile in
polar solvents.
Figure 2B shows the change in vanillyl nonanoate
with time in several solvents containing water at 25 °C.
The half-life periods of vanillyl nonanoate in methanol
and ethanol containing 1% water were about 105 and
90 h, respectively. The decomposition rates of vanillyl
nonanoate in the solvents were little different from
those in only methanol and ethanol, respectively. How-
ever, the decomposition rate of vanillyl nonanoate in
methanol containing 50% water was faster than that
in methanol alone, and the half-life period was ∼45 h.
In contrast, vanillyl nonanoate was stable even in ethyl
acetate containing 1% or saturated (3.1%) water.
Decom p osition P r od u cts of Va n illyl Non a n oa te
w ith Meth a n ol a n d Eth a n ol. Dissolution of vanillyl
nonanoate in methanol and ethanol gave decomposition
products, compounds 9 and 10, respectively. Compound
9 was determined to be C9H12O3 by HRMS measure-
ment. The 1H NMR spectrum of 9 showed three aro-
matic protons (δ 6.85 m) of which the chemical shift
values and signal patterns were very closely similar to
those of vanillyl nonanoate. There were three substitu-
ent groups in the aromatic ring, a hydroxyl group (δ 5.60
s), an oxymethylene group (δ 4.37 s), and a methoxyl
group (δ 3.90 s). An aliphatic methoxyl group was
observed in the spectrum (δ 3.36 s). From these data,
the structure of 9 was established to be 4-hydroxy-3-
methoxybenzyl methyl ether, that is, methyl vanillyl
ether (Figure 1). Compound 10 was determined as
C10H14O3 by HRMS measurement. This molecular for-
mula indicated an addition of CH2 to that of 9. More-
1
over, the H NMR spectrum of 10 was similar to that of
9 except for an ethyl ether group (δ 3.52 q, δ 1.23 t).
From these data, compound 10 was determined to be
4-hydroxy-3-methoxybenzyl ethyl ether, that is, ethyl
vanillyl ether (Figure 1). Both concomitant products
obtained as white insoluble material during the decom-
position of vanillyl nonanoate by methanol and ethanol
were determined to be nonanoic acid by 1H NMR
analysis (data not shown).
Sta bility of Ca p sia te in Or ga n ic Solven ts. Figure
3 shows the change in capsiate (1), a major naturally
occurring capsinoid, with time in ethyl acetate and
methanol at 25 °C. Capsiate was stable in ethyl acetate,
whereas it was labile in methanol, and the half-life
period was ∼50 h. The decomposition product from
capsiate was identified as methyl vanillyl ether by
various spectral data. These results were similar to

Stability of Capsinoid in Various Solvents
J. Agric. Food Chem., Vol. 49, No. 8, 2001 4029
vehicle (Figure 2A). The decomposition rates of vanillyl
nonanoate in methanol, ethanol, and aqueous surfactant
vehicle were the same and were faster than that in
DMSO. Therefore, vanillyl nonanoate probably decom-
poses more easily in a protic than in an aprotic solvent.
High polarity of a solvent is probably responsible for
the decomposition of vanillyl nonanoate because among
the aprotic solvents used, the polarity, namely, the
dielectric constant, of DMSO (45.0) is the highest (9).
The decomposition rate of vanillyl nonanoate in metha-
nol containing 50% water was faster than that in
methanol alone (Figure 2B). This phenomenon can also
be explained by the higher polarity of H2O (dielectric
constant ) 78.5) than that of methanol (32.6). On the
other hand, vanillyl nonanoate was stable even in water-
saturated (3.1%) ethyl acetate. This suggested that
vanillyl nonanoate was protected from an attack of
water by excess ethyl acetate. Interestingly, the decom-
position rate of vanillyl nonanoate in aqueous surfactant
vehicle, an almost aqueous solvent, was little different
from those in only methanol and ethanol. Because
Tween 80, a surface active agent, has a very low critical
micelle concentration (0.012 mM), it is certainly present
as micelles in the aqueous surfactant vehicle. In addi-
tion, vanillyl nonanoate would be incorporated into the
micelles. Therefore, the Tween 80 probably acts as a
shield of vanillyl nonanoate against the attack of the
aqueous solvent. In the case of a general solvolysis of
simple esters, for example, alcoholysis, the resulting
products are a corresponding alkyl carboxylate and an
alcohol. However, the decomposition products of vanillyl
nonanoate induced by methanol were methyl vanillyl
ether and nonanoic acid, whereas ethyl vanillyl ether
and nonanoic acid were induced by ethanol. In brief, the
solvolysis of capsinoid by an alcohol gave a correspond-
ing alkyl ether and a carboxylic acid. Therefore, the
decomposition reaction of capsinoid is different from the
usual solvolysis of simple esters.
The above considerations strongly support the theory
that the decomposition process of capsinoid involves SN1
solvolysis through nucleophilic displacement of the
benzyl cation. The conjectural process of capsinoid
decomposition is as follows: first, the polarization
between the R-methylene of the benzyl moiety and the
neighboring oxygen atom in capsinoid is increased by
the presence of protic and/or polar solvents; then, the
polarized portion is easily cleaved by means of hetero-
lytic cleavage to give a benzyl cation and a carboxylate
anion; finally, the solvent attacks the benzyl cation as
a nucleophilic reagent, in the case of methanol, and
methyl vanillyl ether is formed. In this process, the
formation of the benzyl cation is the rate-determining
step. Hence, the decomposition rate depends on the
stability of the benzyl cation formed. It is natural that
no difference in the decomposition rates between vanil-
lyl nonanoate and capsiate was observed (Figure 3),
because the structural difference in the two compounds
is only in their acyl moieties.
F igu r e 3. Change in capsiate with time in organic solvents
at 25 °C: (O) ethyl acetate; (b) methanol.
F igu r e 4. Change in vanillyl nonanoate analogues with time
in methanol at 25 °C: (b) vanillyl nonanoate; (2) 4-acetoxy-
3-methoxybenzyl nonanoate; (4) 3,4-dimethoxybenzyl nona-
noate; (]) 3-methoxybenzyl nonanoate; (9) p-hydroxybenzyl
nonanoate; (O) m-hydroxybenzyl nonanoate; ([) o-hydroxy-
benzyl nonanoate.
those found for vanillyl nonanoate. Therefore, natural
capsinoid is probably stable in nonpolar solvents and
labile in polar solvents and probably tends to decompose
in protic solvents such as alcohol and water.
Sta bility of Va n illyl Non a n oa te An a logu es in
Or ga n ic Solven ts. The stability of vanillyl nonanoate
analogues in ethyl acetate and methanol was measured
at 25 °C. Figure 4 shows the change in vanillyl nonanoate
and its six analogues with time in methanol. Three
analogues, 4-acetoxy-3-methoxybenzyl nonanoate, o-hy-
droxybenzyl nonanoate, and p-hydroxybenzyl nonanoate,
were labile in methanol, whereas the others were stable
in methanol. The decomposition rates of the three
vanillyl nonanoate analogues were slower than that of
vanillyl nonanoate. The half-life period of p-hydroxy-
benzyl nonanoate (310 h) was ∼3-fold longer than that
of vanillyl nonanoate. The half-life periods of 4-acetoxy-
3-methoxybenzyl nonanoate and o-hydroxybenzyl nona-
noate were not observed within 400 h. On the other
hand, all vanillyl nonanoate analogues were stable in
ethyl acetate (data not shown).
In our preliminary experiment, no decomposition of
benzyl nonanoate occurred in any solvents. Therefore,
it is apparent that the substitution groups in the
benzene ring of capsinoid would greatly contribute to
the formation and stabilization of the benzyl cation. Our
experiments to determine the stability of vanillyl
nonanoate analogues (Figure 4) revealed the following:
(i) a phenolic hydroxyl group is essential for the decom-
position of capsinoid because no or only slight decom-
DISCUSSION
Vanillyl nonanoate was stable in ethyl acetate, chlo-
roform, hexane, and acetone, whereas it was labile in
methanol, ethanol, DMSO, and aqueous surfactant

4030 J. Agric. Food Chem., Vol. 49, No. 8, 2001
Sutoh et al.
dihydrocapsiate, from the fruits of a nonpungent culti-
var, CH-19 Sweet, of pepper (Capsicum annuum L.). J .
Agric. Food Chem. 1998, 46, 1695-1697.
position was observed for nonphenolic analogues such
as 3,4-dimethoxybenzyl nonanoate, 4-acetoxy-3-meth-
oxybenzyl nonanoate, and 3-methoxybenzyl nonanoate;
(ii) from the results for positional isomers, o-, m-, and
p-hydroxybenzyl nonanoate, the effect of the phenolic
hydroxyl group is maximally displayed by its para-
position, followed by the ortho- and meta-positions; (iii)
the methoxyl group of the meta-position in capsinoid
prompts the decomposition because the decomposition
rate of vanillyl nonanoate is faster than that of p-
hydroxybenzyl nonanoate.
These considerations strongly support the hypothesis
that the decomposition of capsinoid depends on the
stability of the formed benzyl cation. In general, the
substitution of an electron-donating group such as a
hydroxyl group on the benzene ring disperses the
positive charge at the R-methylene of the benzyl moiety,
consequently stabilizing the benzyl cation. The hydroxyl
group in a phenol is an ortho-para directing group, so
that the benzyl cations of o- and p-benzyl phenols are
highly stabilized.
In conclusion, capsinoid was stable in nonpolar sol-
vents, whereas it was labile in polar solvents and
especially tended to decompose in protic solvents such
as alcohol and water. The decomposition proceeds by
SN1 solvolysis via the formed benzyl cation, and the
p-hydroxyl group in the benzyl moiety of capsinoid
contributes to the formation and stabilization of the
benzyl cation. It is necessary for further use of capsinoid
to prevent its decomposition, that is, to prevent the
generation of the benzyl cation. Protection by a nonpolar
or low-polar solvent and a detergent from the attack of
active solvents could be a useful way to achieve stabi-
lization of capsinoid in solvents.
(2) Kobata, K.; Sutoh, K.; Todo, T.; Yazawa, S.; Iwai, K.;
Watanabe, T. Nordihydrocapsiate, a new capsinoid from
the fruits of a nonpungent pepper, Capsicum annuum.
J . Nat. Prod. 1999, 62, 335-336.
(3) Nelson, E. K.; Dawson, L. E. The constitution of cap-
saicin, the pungent principle of Capsicum. III. J . Am.
Chem. Soc. 1923, 45, 2179-2181.
(4) Kosuge, S.; Inagaki, Y.; Okumura, H. Studies on the
pungent principles of red pepper. Part VIII. On the
chemical constitutions of the pungent principles. (5) On
the chemical constitution of the pungent principle II.
Nippon Nogei Kagaku Kaishi 1961, 35, 923-927.
(5) Buck, S. H.; Burks, T. F. The neuropharmacology of
capsaicin: Review of some recent observations. Phar-
macol. Rev. 1986, 38, 179-226.
(6) Govindarajan, V. S.; Sathyanarayana, M. N. Capsicums
Production, technology, chemistry, and quality. Part V.
Impact on physiology, pharmacology, nutrition, and
metabolism; Structure, pungency, pain, and desensitiza-
tion sequences. CRC Crit. Rev. Food Sci. Nutr. 1991,
29, 435-474.
(7) Yazawa, S.; Niwa, S.; Watanabe, T.; Fushiki, T. Bio-
synthesis of capsaicinoid-like substances in Capsicum
fruits and their physical activities in the human body.
Suppl. J . J pn. Soc. Hortic. Sci. 1998, 67, 296.
(8) Walpole, C. S. J .; Wrigglesworth, R.; Bevan, S.; Camp-
bell, E. A.; Dray, A.; J ames, I. F.; Masdin, K. J .; Perkins,
M. N.; Winter, J . Analogues of capsaicin with agonist
activity as novel analgesic agents; Structure-activity
studies. 2. The amide bond “B-region”. J . Med. Chem.
1993, 36, 2373-2380.
(9) Hendrickson, L. C.; Cram, A. C.; Hammond, G. S.
Nucleophilic substitution at saturated carbon. In Or-
ganic Chemistry, 3rd ed.; McGraw-Hill: Kougakusha,
Tokyo, J apan, 1970; p 388.
ACKNOWLEDGMENT
We thank Yasufumi Morita of the Toyama National
College of Technology for measuring the HRMS data.
Received for review March 13, 2001. Revised manuscript
received J une 1, 2001. Accepted J une 1, 2001. This work was
supported by a grant-in-aid for science research (B) from The
J apan Society for the Promotion of Science.
LITERATURE CITED
(1) Kobata, K.; Todo, T.; Yazawa, S.; Iwai, K.; Watanabe,
T. Novel capsaicinoid-like substances, capsiate and
J F0103424