Capsaicinol: Synthesis by allylic oxidation and its effect on TRPV1-expressing cells and adrenaline secretion in rats

Article abstract of DOI:10.1271/bbb.60064

Capsaicinol is an ingredient of hot red pepper. In this study, we developed a novel method for capsaicinol synthesis and examined capsaicinol's physiological effects on capsaicin receptor (TRPV1)-related actions. Allylic oxidation of capsaicin by palladiu

Full text of DOI:10.1271/bbb.60064

Biosci. Biotechnol. Biochem., 70 (8), 1904–1912, 2006
Capsaicinol: Synthesis by Allylic Oxidation and Its Effect
on TRPV1-Expressing Cells and Adrenaline Secretion in Rats
Kenji KOBATA,1 Takahito IWASAWA,1 Yusaku IWASAKI,1 Akihito MORITA,1
Yuichi SUZUKI,1 Hiroe KIKUZAKI,2 Nobuji NAKATANI,2 and Tatsuo WATANABE
y
1;
1Graduate School of Nutritional and Environmental Sciences and COE program in the 21st century,
University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka 422-8526, Japan
2Graduate School of Human Life Science, Osaka City University, 3-3-138 Sugimoto,
Sumiyoshi-ku, Osaka 558-8585, Japan
Received February 3, 2006; Accepted April 13, 2006; Online Publication, August 23, 2006
[doi:10.1271/bbb.60064]
Capsaicinol is an ingredient of hot red pepper. In this
study, we developed a novel method for capsaicinol
synthesis and examined capsaicinol’s physiological ef-
fects on capsaicin receptor (TRPV1)-related actions.
Allylic oxidation of capsaicin by palladium acetate
(Pd(OAc)2) resulted in the formation of (ꢀ)-capsaicinol
acetate at a 7.2% yield in a single step. The effectiveness
of (ꢀ)-capsaicinol in TRPV1 activation (EC50 ¼ 1:1 ꢀM)
a member of the transient receptor potential vanilloid
(TRPV) subfamily, has been cloned and characterized
in detail.6–8) Activation of TRPV1 is responsible for a
majority of capsaicin effects, including the burning
sensation. TRPV1 is a non-selective cation channel
activated by noxious heat (> 43 ꢀC), protons, and
vanilloids. Vanilloids are a group of TRPV1 agonists
whose structures contain a 4-hydroxy-3-methoxybenzyl
moiety. Several naturally occurring and chemically
synthesized vanilloids have been reported.9–13)
Capsaicin stimulates sensory neurons to increase
catecholamine secretion via activation of the sympa-
thetic nervous system.14) An enhancement of energy
metabolism due to catecholamine secretion results in
thermogenesis, and consequently suppresses adipose
tissue deposition.15,16) Although capsaicin is an attrac-
tive candidate for treating obesity, its strong pungency
and harmful stimuli limit its usage as a food additive and
in drugs. Several non-pungent capsaicin analogs with
longer acyl chains than capsaicin have a potential
similar to that of capsaicin for the enhancement of
adrenaline secretion in rats17,18) and for promoting
energy metabolism.19) Previously, we observed that
catecholamine secretion in the presence of capsaicin
and its analogs was due to the activation of TRPV1 in
the sensory neurons.18) Furthermore, a naturally occur-
ring and non-pungent capsaicin analog, capsiate, in-
creased adrenaline secretion and suppressed body fat
accumulation in mice.20,21) Capsiate also possessed a
TRPV1 activation potential comparable to that of
capsaicin.13)
was found to be weaker than that of capsaicin (EC50
¼
0:017 ꢀM), whereas the efficacy of (ꢀ)-capsaicinol
reached 75% of that of capsaicin. Intravenous admin-
istration of (ꢀ)-capsaicinol in anesthetized rats dose-
dependently enhanced adrenaline secretion from the
adrenal gland. The response to a 5 mg/kg-dose of (ꢀ)-
capsaicinol was comparable to that of a 0.05 mg/kg-dose
of capsaicin. The relative pungency of capsaicinol to
capsaicin was coincident with the relative effectiveness
in inducing these TRPV1-related actions.
Key words: capsaicinol; allylic oxidation; transient re-
ceptor potential vanilloid 1 (TRPV1); adren-
aline secretion; pungency
Capsaicinol, N-(4-hydroxy-3-methoxybenzyl)-7-hy-
droxy-8-methyl-5E-nonenamide, is an antioxidant pres-
ent in the fruits of the pepper Capsicum frutescens L.1–3)
The antioxidant activity of capsaicinol against linoleic
acid oxidation is comparable to that of ꢀ-tocopherol and
capsaicin, a well-known pungent ingredient of hot
pepper.1) Capsaicinol is less pungent than capsaicin
although its chemical structure is very similar to that of
capsaicin (Fig. 1).
In the present study, we attempted to synthesize
capsaicinol by a novel method, viz., allylic oxidation of
capsaicin by palladium(II) acetate (Pd(OAc)2). This
synthesis was performed prior to physiological assays
Capsaicin not only induces a burning sensation, but
also elicits various physiological and pharmacological
responses.4,5) TRPV1, a specific capsaicin receptor and
y
To whom correspondence should be addressed. Tel: +81-54-264-5543; Fax: +81-54-264-5550; E-mail: watanbt@u-shizuoka-ken.ac.jp
Abbreviations: TRPV1, transient receptor potential vanilloid 1; HEK 293 cells, human embryonic kidney 293 cells; ODV, oil dilution value; SHU,
Scoville heat units

Synthesis and TRPV1-Related Actions of Capsaicinol
1905
since capsaicinol could not be obtained in sufficient
quantities from natural sources. Next, in order to study
whether capsaicinol activates TRPV1, we carried out
electrophysiological analysis using the whole cell patch
clamp method on recombinant rat TRPV1 that was
stably expressed in human embryonic kidney (HEK) 293
cells. Further, we investigated the effect of capsaicinol
on adrenaline secretion from the adrenal gland by in situ
experiments in anesthetized rats. Finally, we discuss the
relationship between the physiological responses, in-
cluding pungency as evaluated by organoleptic tests, and
the lipophilicity of capsaicinol.
instrument at 399.65 and 100.40 MHz respectively. IR
spectra were recorded on a Jasco (Tokyo) FT/IR-550
spectrophotometer, and UV spectra were recorded on a
Jasco UVIDEC 660 spectrophotometer. HR-FABMS
measurements were carried out on a JEOL JMS-700
spectrometer.
Compound 1: colorless oil; HR-FABMS m=z
386.1929 (calcd. for C20H29NO5Na 386.1944); IR (film)
ꢁmax 3300, 2930, 2860, 1720, 1640, 1520, 1430, 1370,
1240, 1160, 1130, 1030, 940, 820, 750 cmꢃ1; UV
1
(MeOH) ꢂmax (") 281 (2700), 229 (6800) nm; H-NMR
(CDCl3) ꢃ 6.86 (1H, d, J ¼ 8:0 Hz, H-50), 6.80 (1H, d,
J ¼ 2:0 Hz, H-20), 6.76 (1H, dd, J ¼ 8:0, 2.0 Hz, H-60),
5.65 (2H, b, OH and NH), 5.45 (1H, dt, J ¼ 9:0, 6.8 Hz,
H-6), 5.08 (1H, d, J ¼ 9:0 Hz, H-7), 4.35 (2H, d, J ¼
6:0 Hz, H-70), 3.89 (3H, s, OCH3), 2.19 (2H, t, J ¼ 7:6
Hz, H-2), 2.00 (3H, s, OAc), 1.72 (3H, s, H-9 or H-10),
1.70 (3H, s, H-9 or H-10), 1.65 (4H, m, H-3 and H-5),
1.30 (1H, m, H-4); 13C-NMR (CDCl3) ꢃ 172.5 (C-1),
170.5 (O–(C=O)–CH3), 146.7 (C-30), 145.2 (C-40),
137.4 (C-8), 130.4 (C-10), 123.7 (C-7), 120.8 (C-60),
114.4 (C-50), 110.8 (C-20), 71.4 (C-6), 56.0 (OCH3),
43.6 (C-70), 36.6 (C-2), 34.7 (C-5), 25.8 (C-9), 25.5
(C-3), 24.9 (C-4), 21.3 (O–(C=O)–CH3), 18.5 (C-10).
Compound 2 ((ꢁ)-capsaicinol-7-O-acetate): colorless
oil; HR-FABMS m=z 386.1942 (calcd. for C20H29-
NO5Na 386.1944); IR (film) ꢁmax 3310, 2960, 2860,
1730, 1646, 1520, 1460, 1430, 1370, 1240, 1160, 1130,
1030, 970, 820 cmꢃ1; UV (MeOH) ꢂmax (") 281 (1800),
Materials and Methods
Chemicals. Capsaicin (> 97% pure), capsazepine
(> 99% pure), and the antibiotic G418 were purchased
from Sigma (MO, USA). Palladium(II) acetate
(Pd(OAc)2, > 97% pure) was purchased from Wako
Pure Chemical (Osaka, Japan). First-grade reagent
capsaicin, containing both capsaicin (60%) and dihy-
drocapsaicin (40%), was purchased from Wako for
the preparation of capsaicinol and its analogs. The
other reagents used in this study were of guaranteed
grade.
Synthesis of (ꢁ)-capsaicinol by allylic oxidation of
capsaicin. The following reagents were added to 200 ml
acetic acid solution: 500 mg of first-grade reagent
capsaicin (containing 300 mg capsaicin) and 500 mg of
Pd(OAc)2. The mixture was then agitated at room
temperature for 15 h. The reaction products were ex-
tracted by the addition of ethyl acetate (250 ml ꢂ 3),
and the ethyl acetate solution was washed with water;
the products were then dried with Na2SO4. The residue
(780 mg) obtained after evaporation was applied to
an MPLC system (Yamazen, Osaka, Japan) under the
following conditions: column, Ultrapack SI-40C, 37
mm i.d. ꢂ 300 mm (Yamazen); eluent, a mixture of
hexane and ethyl acetate (EtOAc); flow rate, 20–30 ml/
min; detection, UV 280 nm. The hexane/EtOAc (60/40)
elution fraction (55 mg) was passed through an alumina
column (20 mm i.d. ꢂ 35 mm) with methanol to elim-
inate highly polar materials. The methanol eluent was
chromatographed to afford 1 (9.7 mg, 2.7% isolated
yield) and a mixture of 2 and 3 using a preparative
HPLC system under the following conditions: column,
J’sphere ODS-H80, 20 mm i.d. ꢂ 150 mm (YMC,
Kyoto, Japan); eluent, 55% methanol (MeOH); flow
rate, 10 ml/min; detection, fluorescence (ex 280 nm, em
320 nm). Compounds 2 (4.2 mg, 1.2% isolated yield)
and 3 (4.8 mg, 1.3% isolated yield) were purified from
their mixture using an HPLC system under the following
conditions: column, Fluofix, 4.6 mm i.d. ꢂ 250 mm
(Wako); eluent, 35% MeOH; flow rate, 1.2 ml/min;
detection, fluorescence (ex 280 nm, em 320 nm). The
1H- and 13C-NMR spectra (TMS as the internal stand-
ard) were recorded on a JEOL (Tokyo, Japan) ꢀ-400
1
230 (4700) nm; H-NMR (CDCl3) ꢃ 6.86 (1H, d, J ¼
8:0 Hz, H-50), 6.81 (1H, d, J ¼ 1:5 Hz, H-20), 6.76 (1H,
dd, J ¼ 8:0, 1.5 Hz, H-60), 5.74 (1H, b, NH), 5.60 (1H,
dt, J ¼ 15:6, 6.8 Hz, H-5), 5.32 (1H, dd, J ¼ 15:6, 6.8
Hz, H-6), 4.90 (1H, t, J ¼ 6:8 Hz, H-7), 4.35 (2H, d,
J ¼ 5:6 Hz, H-70), 3.88 (3H, s, OCH3), 2.17 (2H, t, J ¼
7:6 Hz, H-2), 2.08 (2H, q, J ¼ 6:8 Hz, H-4), 2.01 (3H, s,
OAc), 1.7 (3H, m, H-3 and H-8), 0.88 (3H, d, J ¼
6:8 Hz, H-9 or H-10), 0.86 (3H, d, J ¼ 6:8 Hz, H-9 or
H-10); 13C-NMR (CDCl3) ꢃ 172.5 (C-1), 170.5 (O–
(C=O)–CH3), 146.7 (C-30), 145.2 (C-40), 133.6 (C-5),
130.5 (C-10), 127.9 (C-6), 120.9 (C-60), 114.4 (C-50),
110.9 (C-20), 79.7 (C-7), 56.0 (OCH3), 43.6 (C-70), 35.8
(C-2), 32.0 (C-8), 31.5 (C-4), 24.9 (C-3), 21.3 (O–
(C=O)–CH3), 18.1 (C-9), 18.1 (C-10).
Compound 3: colorless oil; HR-FABMS m=z
386.1921 (calcd. for C20H29NO5Na 386.1944); IR (film)
ꢁmax 3310, 2960, 1720, 1646, 1520, 1460, 1430, 1370,
1240, 1160, 1130, 1030, 970 cmꢃ1; UV (MeOH) ꢂmax
(") 281 (1900), 230 (5300) nm; 1H-NMR (CDCl3) ꢃ 6.86
(1H, d, J ¼ 8:0 Hz, H-50), 6.81 (1H, d, J ¼ 1:5 Hz, H-
20), 6.76 (1H, dd, J ¼ 8:0, 1.5 Hz, H-60), 5.74 (1H, b,
NH), 5.66 (1H, dd, J ¼ 15:6, 6.8 Hz, H-7), 5.31 (1H, dd,
J ¼ 15:6, 7.4 Hz, H-6), 5.19 (1H, q, J ¼ 7:4 Hz, H-5),
4.35 (2H, d, J ¼ 5:6 Hz, H-70), 3.88 (3H, s, OCH3), 2.25
(1H, m, H-8), 2.21 (2H, t, J ¼ 7:6 Hz, H-2), 2.02 (3H, s,
OAc), 1.65 (4H, m, H-3 and H-4), 0.97 (3H, d, J ¼
6:8 Hz, H-9), 0.97 (3H, d, J ¼ 6:8 Hz, H-10); 13C-NMR
(CDCl3) ꢃ 172.3 (C-1), 170.5 (O–(C=O)–CH3), 146.7

1906
K. KOBATA et al.
(C-30), 145.3 (C-40), 141.5 (C-7), 130.4 (C-10), 125.0 (C-
6), 120.9 (C-60), 114.5 (C-50), 110.8 (C-20), 74.4 (C-5),
56.0 (OCH3), 43.7 (C-70), 36.2 (C-2), 34.1 (C-4), 30.7
(C-8), 22.1 (C-9), 22.1 (C-10), 21.4 (C-3), 21.3 (O–
(C=O)–CH3).
A solution of 1 N NaOH was added to the methanol
solution of compound 2, and the mixture was maintained
at 50 ꢀC for 2 h. After neutralization with 1 N HCl, the
mixture was treated with EtOAc to afford (ꢁ)-capsai-
cinol. The spectral data of this synthesized (ꢁ)-capsai-
cinol were identical to those of the authentic sample.1)
Two enantiomers, (+)- and (ꢃ)-capsaicinol, were
obtained on repeated HPLC purification from the
racemic mixture using a chiral column. The HPLC
conditions are described in the legend to Fig. 4.
parameter estimation were carried out using GraphPad
Prism 4 (Graph Pad Software, CA, USA).
Measurement of adrenaline secretion in rats. Adrena-
line secretion in response to the chemicals was measured
from the adrenal vein of anesthetized rats by the use of a
continuous blood sampling system, as described pre-
viously.18) Briefly, a rat anesthetized with ꢀ-chloralose-
urethane was placed on a heating pad to maintain the
body temperature at approximately 37 ꢀC. Two cathe-
ters, one each for sample injection and blood sampling,
were inserted into the femoral vein and the adrenal vein
respectively. Heparinized saline was transfused into the
femoral vein during the experiment. Adrenal venous
blood was collected after injection of the sample
solution (2 vol % ethanol, 10 vol % Tween 80, and
88 vol % saline, volume injected: 1 ml/kg). TRPV1
inhibitor capsazepine was injected 2 min before sample
injection. As an internal standard, 3,4-dihydroxybenzyl-
amine was added to the collected blood. After centri-
fugation, the plasma catecholamines were purified with
activated alumina and measured by HPLC-EC. Animal
experiment procedures met the guidelines for the care
and use of laboratory animals of the University of
Shizuoka.
Optimization of capsaicinol synthesis. To obtain
optimal reaction conditions for (ꢁ)-capsaicinol acetate
synthesis, small-scale reactions were performed as
follows: To 1 ml of the acetic acid solution of 10 mM
capsaicin (> 97% pure), the following were added: 0.1–
10 mol fraction of the metal acetates, viz., palladium(II)
acetate (Pd(OAc)2), mercury(II) acetate (Hg(OAc)2),
lead(IV) acetate (Pb(OAc)4), and manganese(III) acetate
(Mn(OAc)3). The reaction mixtures were agitated at
25 ꢀC or 70 ꢀC. Part of the reaction mixture was then
subjected to HPLC analysis to quantify the products and
remaining capsaicin. The HPLC conditions are describ-
ed in the legend to Fig. 2. Authentic compounds for
quantification were prepared using the large-scale
reaction described above.
Evaluation of capsaicinol pungency
Oil dilution method. The oil dilution method was
carried out as described by Watanabe et al.,17) with
slight modifications. The test compound was diluted
with olive oil to obtain three dilutions that were near to
the threshold of pungency predicted by a preliminary
test. Twenty microliters of the solution was dropped on
the tongues of five trained panels. A linear graph was
plotted based on the degree of dilution versus the
percentage of panels who felt the pungency. The value
of the degree of dilution that corresponded with 50% of
the panels who felt the pungency, i.e., the oil dilution
value (ODV), was defined as the threshold of pungency
for the test compound.
Scoville heat units. Scoville heat units (SHU) of (ꢁ)-
capsaicinol were measured using the ISO method.23) An
ethanol solution of (ꢁ)-capsaicinol (5 mg/ml) was
diluted with 5% sucrose aqueous solution to obtain five
dilutions that were near the threshold of pungency
predicted by a preliminary test. Five milliliters of the
solution was tasted for 30 s, but not ingested. The SHU
of (ꢁ)-capsaicinol was calculated by the same method
as the oil dilution values.
Current recording by whole cell patch clamp method.
HEK 293 cells stably expressing rat TRPV1 were
obtained as described previously,22) and were main-
tained in Dulbecco’s modified Eagle’s medium supple-
mented with 10% fetal bovine serum, 100 units/ml
penicillin, 100 mg/ml streptomycin, 0.25 mg/ml ampho-
tericin B, and 60 mg/ml G418 under 5% CO2 at 37 ꢀC.
The HEK 293 cells with or without TRPV1 expres-
sion were spread on glass coverslips coated with poly-L-
lysine, and were cultured in the culture medium 1–3 d
prior to the assay. The glass coverslips with the cells
were transferred to a small chamber on the stage of
an inverted microscope (IMT2, Olympus, Tokyo). The
chamber was perfused continuously with bath solution
containing 135 mM NaCl, 2 mM KCl, 2 mM MgSO4,
10 mM glucose, 5 mM EGTA, and 10 mM HEPES (pH
7.4). All recordings were performed using standard
whole cell patch clamp methods. Glass pipette resistance
ranged from 3 to 5 Mꢀ. The pipette solution contained
140 mM KCl, 5 mM EGTA, and 10 mM HEPES (pH 7.4).
The test compound was dissolved in the bath solution
containing 0.1% dimethyl sulfoxide. A patch clamp
amplifier (CEZ 2300, Nihon Kohden, Tokyo) was used
to record the membrane currents. The intracellular
voltage was clamped at minus 50 mV. All recordings
were made at room temperature. Curve fitting and
These protocols were designed in accordance with the
principles of the Declaration of Helsinki of the World
Medical Association.
Measurement of capsaicinol lipophilicity. The lipo-
philicity of (ꢁ)-capsaicinol was evaluated as the
partition coefficient according to the n-octanol/water
flask shaking method.24) One milliliter of n-octanol
solution of (ꢁ)-capsaicinol (10 mM) was added to 1 ml

Synthesis and TRPV1-Related Actions of Capsaicinol
1907
Fig. 1. Schematic Representation of (ꢁ)-Capsaicinol Synthesis via Reaction Products (1–3) Generated by Palladium(II) Acetate (Pd(OAc)2)
Oxidation of Capsaicin.
water in a 5-ml screw vial. The mixture was agitated
vigorously at 20 ꢀC for 60 min. After centrifugation at
1;600 ꢂ g for 1 min, the (ꢁ)-capsaicinol concentrations
of the n-octanol and water fractions were analyzed
by the HPLC system, as described previously.13) The
lipophilicity was represented as log P, i.e., the logarithm
of the solute concentration of the n-octanol fraction
divided by that of the water fraction.
at the end of the side chain of 1 shifted by approximately
1 ppm higher field than those in capsaicin, and both the
signal patterns were singlets. Furthermore, one of the
olefinic carbons (ꢃ 137.4) was indicated to be quaternary
by DEPT analysis of the 13C-NMR of 1. Therefore, the
olefin in 1 was considered to be at the 7 position. A
proton of a methine moiety bound to the acetoxy group
was observed to be a double triplet at ꢃ 5.45. At 9.0 and
6.8 Hz, it coupled with 7-olefinic methine at ꢃ 5.08 and a
5-methylene group at ꢃ 1.65 respectively. Based on all
these data, the structure of 1 was confirmed to be N-(4-
hydroxy-3-methoxybenzyl)-6-acetoxy-8-methyl-7-non-
enamide.
Results and Discussion
Synthesis of (ꢁ)-capsaicinol by palladium acetate
oxidation of capsaicin
Capsaicinol structurally resembles capsaicin, and they
differ only with respect to their side chains. In the case
of capsaicinol, an additional hydroxy group is attached
at the 7 position. Further, an olefin is located at the
5 position, and not at the 6 position as in capsaicin
(Fig. 1). In brief, rearrangement of the olefin and
subsequent hydroxylation at the allylic position convert
capsaicin to capsaicinol. The oxidation of olefins by
palladium(II) acetate (Pd(OAc)2) yields a variety of
products, including allylic acetates with the rearranged
olefin.25) Therefore, we expected that Pd(OAc)2 might
induce capsaicinol production from capsaicin in a few
steps.
In the 1H-NMR spectrum of 2, a methine proton
bound to the acetoxy group was observed to be a triplet
at ꢃ 4.90. At 6.8 Hz, it coupled with an isopropyl
methine at ꢃ 1.7 and an olefinic methine at ꢃ 5.32, as
determined by HH COSY analysis. At 15.6 Hz, the
olefinic methine further coupled with another olefinic
methine at ꢃ 5.60. These data indicated the existence of
1
a 7-acetoxy-8-methyl-5E-nonene moiety in 2. The H-
and 13C-NMR spectra of 2 were very similar to those of
capsaicinol, except for acetoxy group-related signals.
Therefore, the structure of 2 was determined to be N-
(4-hydroxy-3-methoxybenzyl)-7-acetoxy-8-methyl-5E-
nonenamide, that is, capsaicinol-7-O-acetate. Hydrolysis
of 2 by an alkali yielded capsaicinol quantitatively.
HPLC analysis revealed that this synthetic capsaicinol
was a racemic mixture (Fig. 4A).
The allylic oxidation of capsaicin was carried out in a
simple manner, i.e., by mixing capsaicin with Pd(OAc)2
in acetic acid at room temperature. HPLC analysis
during the course of the reaction indicated the gener-
ation of three main products (1–3) (Fig. 2A). These
compounds were purified by repeated chromatographic
separations. The molecular formulae of 1–3 were
deduced to be C20H29NO5 by high resolution fast atom
bombardment mass spectroscopy (HR-FABMS) meas-
urement, and this indicated that C2H2O2 had been added
to capsaicin. Their 1H-NMR spectra showed a single
peak of three protons at ꢃ 2.0, and their 13C-NMR
spectra showed a carbonyl carbon of the ester group at
ꢃ 170.5 and a methyl group at ꢃ 21.3. These data
suggested that compounds 1–3 were the acetoxy group
adducts of capsaicin.
1
The H-NMR spectrum of 3 revealed a proton of an
acetoxy methine at ꢃ 5.19 and the coupling of two
olefinic protons, resulting in trans configuration (ꢃ 5.66
and ꢃ 5.31, 15.6 Hz). Although these signals were
similar to those of 2, the acetoxy methine and the
olefinic methine at ꢃ 5.66 were coupled with aliphatic
methylene and isopropyl methine respectively. There-
fore, the structure of 3 was confirmed to be N-(4-
hydroxy-3-methoxybenzyl)-5-acetoxy-8-methyl-6E-non-
enamide.
In general, the formation of allylic acetate with
simultaneous olefin rearrangement by Pd(OAc)2 is
initiated by the addition of an acetoxy palladium ion
(þPdOAc) and an acetoxy ion (ꢃOAc) to olefin; this is
In the 1H-NMR spectrum of 1, the two methyl groups

1908
K. KOBATA et al.
(2), we performed small-scale experiments to optimize
the reaction conditions, such as temperature, ratio of
starting materials, choice of other metal acetates, and
solvents used. The low reaction temperature and low
mole fraction of the metal acetate yielded only a trace
amount of products. Among several metal acetates,
Pd(OAc)2 and mercury(II) acetate (Hg(OAc)2) yielded
products. However, the potential of Hg(OAc)2 to yield
these products was lower than that of Pd(OAc)2. No
product similar to 1–3 was formed on reacting capsaicin
with lead(IV) acetate (Pb(OAc)4) or manganese(III)
acetate (Mn(OAc)3), in spite of consumption of cap-
saicin during the reaction. A higher mole fraction of
the metal acetate only increased capsaicin consumption,
and no product was formed. These results led to the
optimization of conditions as follows: 10 mM capsaicin
and an equivalent concentration of Pd(OAc)2 in acetic
acid at 70 ꢀC for 20 h or more. Figure 2B shows the time
dependency of product (1–3) yields and the remaining
capsaicin under these conditions. The capsaicin concen-
tration declined drastically, and the capsaicin was
completely consumed before 20 h. On the other hand,
the concentrations of 2 and 3 increased gradually, and
the yield of 2 reached 7.2% after 20 h. A preference for
the formation of 1 was indicated by time course analysis
of the products, particularly at the early reaction time
(Fig. 2B), but the long reaction time caused a gradual
reduction in yield. This was probably due to further
oxidation of 1 by Pd(OAc)2.
The rate of production of 1–3 to the consumption of
capsaicin did not exceed 25% at any point of time. In the
case of simple olefins, metal acetates yield high amounts
of the corresponding allylic acetate.25–27) The functional
groups of capsaicin, such as amide and the phenolic
hydroxy group, probably resulted in unsuitable reactions
with the metal acetates.
Since capsaicinol cannot be obtained in sufficient
quantities from natural sources, the development of a
simple method for capsaicinol synthesis is beneficial in
study of the biological and physiological activities of
capsaicinol. Previous work demonstrated the synthesis
of (ꢁ)-capsaicinol from ꢃ-valerolactone in nine steps,
and the overall yield was 18%.2) In the present paper,
we have proposed a novel method of (ꢁ)-capsaicinol
synthesis. This method is very useful because it is
simple and easy to perform, it is not time consuming,
and does not require the use of complicated or expensive
devices. The only disadvantage is the low product yield.
Further experiments to improve the product yield are in
progress.
Fig. 2. Qualitative and Quantitative Analyses of the Reaction
Products (1–3) by Pd(OAc)2 Oxidation of Capsaicin.
A, The HPLC chromatogram of the reaction mixture. HPLC
conditions were as follows: column, Inertsil Ph 4.6 mm i.d. ꢂ
250 mm (GL Sciences, Tokyo); solvent, 35% acetonitrile; flow rate,
1 ml/min; detection, fluorescence, ex 280 nm, em 320 nm. B, Time
dependency of the yields of 1–3 and remaining capsaicin on the
reaction conditions: 10 mM capsaicin and 10 mM Pd(OAc)2 in acetic
acid at 70 ꢀC.
followed by the formation of an alternative olefin due to
the elimination of HPdOAc. Harris et al. reported that
the addition of the two ions to the olefin tends to follow
Markovnikov’s rule, and that allylic acetate is formed in
preference to enol acetate on elimination of HPdOAc.25)
Our results with regards to the formation and ratios of 1
and 2 corresponded with Markovnikov’s rule. However,
the reaction mechanism for the formation of product 3,
which was unexpected, is not clear.
The allylic oxidation of capsaicin by Pd(OAc)2
yielded the 7-O-acetate of (ꢁ)-capsaicinol along with
its isomers in a single step. The low isolation yields
of these products might be a result of the low reaction
yield and loss of product during the purification process.
TRPV1 activation potential of capsaicinol
Capsaicinol belongs to a class of capsaicin analogs
that are less pungent. We carried out an electrophysio-
logical analysis to evaluate the effectiveness of cap-
saicinol in TRPV1 activation using the whole cell patch
clamp method. (ꢁ)-Capsaicinol evoked an inward
current when the intracellular voltage was clamped at
Optimization of capsaicinol synthesis
To improve the reaction yield of capsaicinol acetate

Synthesis and TRPV1-Related Actions of Capsaicinol
1909
Fig. 3. Electrophysiological Analysis of (ꢁ)-Capsaicinol (CL) and Capsaicin (CAP).
A, A trace of the current recorded on 10 mM CAP and (ꢁ)-CL perfusion (solid bar) in HEK 293 cells with (upper) or without (lower) TRPV1
expression. B, The dose-response curve of (ꢁ)-CL (solid circle) and CAP (solid square). The responses were normalized by the intensity of the
current recorded by 1 mM CAP perfusion. Values are represented as mean ꢁ SEM (n ¼ 6). C, A trace of the current recorded on 1 mM CAP and
20 mM capsazepine (CPZ) perfusion (upper), and 10 mM (ꢁ)-CL and 20 mM CPZ perfusion (lower). D, The normalized responses of 1 mM CAP and
10 mM (ꢁ)-CL with or without 20 mM CPZ. Values are represented as mean ꢁ SEM (n ¼ 9). ꢄP < 0:01, unpaired t test.
minus 50 mV in TRPV1-expressing HEK 293 cells,
while no current was observed in non-transfected cells
(Fig. 3A). The capsaicinol-evoked inward current ob-
served during the perfusion of capsaicinol disappeared
immediately after the capsaicinol was washed out; this
decline of capsaicin-evoked current was gradual. The
response of (ꢁ)-capsaicinol increased dose-dependent-
ly, and the EC50 and the Hill slope were 1.1 mM and 1.8
respectively (Fig. 3B). The potency of (ꢁ)-capsaicinol
was 60 times weaker than that of capsaicin (EC50
¼
0:017 mM, Hill slope = 1.3). Maximum current was
recorded when the concentration of (ꢁ)-capsaicinol
was 10 mM (minus 1:25 nA ꢁ 0:48, mean ꢁ S.D., n ¼
6), and the value obtained was approximately 75% of
that obtained with 1 mM capsaicin (minus 1:66 nA ꢁ
0:34, n ¼ 6). The currents caused by 10 mM (ꢁ)-
capsaicinol were completely eliminated by simultane-
ous perfusion with 20 mM capsazepine, a competitive
TRPV1 antagonist (Fig. 3C, D). The capsaicin-evoked
inward current was also significantly attenuated by
capsazepine.
Capsaicinol has an asymmetric carbon at its 7 position
(Fig. 1). The absolute configuration of the 7 position of
naturally occurring capsaicinol is R, and its optical
rotation has a negative value,3) i.e., (ꢃ)-capsaicinol.
Analysis of chemically prepared capsaicinol by an
HPLC system equipped with a chiral column showed
two peaks of the same area (Fig. 4A). Therefore, the
chemically prepared capsaicinol was probably a racemic
Fig. 4. Optical Resolution of (ꢁ)-Capsaicinol and the TRPV1
Activation Potential of the Enantiomers.
A, Chromatogram of chemically synthesized (ꢁ)-capsaicinol
using a HPLC system under the following conditions: column,
Chiracell OJ-H 4.6 mm i.d. ꢂ 150 mm (Daicel Chemical, Osaka,
Japan); solvent, hexane/IPA = 70/30; flow rate, 1 ml/min; detec-
tion, UV 280 nm. B, The TRPV1 responses of 10 mM racemic
mixture and two enantiomers of capsaicinol, A and B, were
normalized by that of 1 mM capsaicin. n = 3.

1910
K. KOBATA et al.
Fig. 5. The Time Course of Adrenaline Secretion in Rats by Intravenous Injection of (ꢁ)-Capsaicinol.
A, Dose dependency of (ꢁ)-capsaicinol, 0 mg/kg (vehicle, open diamond), 0.5 mg/kg (1.6 mmol/kg, closed square), 2.5 mg/kg (8.0 mmol/kg,
closed triangle) 5.0 mg/kg (16 mmol/kg, closed circle), and 0.05 mg/kg capsaicin (0.16 mmol/kg, open square). B, Effect of coinjection of
20 mg/kg capsazepine (53 mmol/kg, closed triangle) with 2.5 mg/kg (ꢁ)-capsaicinol. The open diamonds and closed triangles show 0 and
2.5 mg/kg (ꢁ)-capsaicinol respectively. Values are represented as mean ꢁ SEM (n ¼ 5{6). ꢄP < 0:05, #P < 0:01, unpaired t test.
mixture, (ꢁ)-capsaicinol; this was confirmed by optical
resolution analysis. There was no difference in TRPV1
activation potency between the enantiomers and the
racemic mixture of capsaicinol (Fig. 4B).
Enhancement of adrenaline secretion in rats
Enhancement of catecholamine secretion prompts
the activation of energy metabolism in mammals. We
reported previously that adrenaline secretion in rats was
a result of TRPV1 activation by intravenously injected
capsaicin.18) Hence, we investigated whether the TRPV1
activator (ꢁ)-capsaicinol induces adrenaline secretion in
rats.
Here, we demonstrated that capsaicinol possesses
the ability to activate TRPV1. Although the potency
of capsaicinol was lower than that of capsaicin or
capsiate,13) its potency was comparable with other
natural vanilloids such as gingerol (EC50 ¼ 0:56 mM),
zingerone (EC50 ¼ 74 mM), and eugenol (EC50 ꢅ 1
mM).9,10) It is well-known that the length of the acyl
moiety in vanilloids is important for TRPV1-related
effects such as pungency, antinociception, and anti-
inflammation,28,29) but except for one report, the effect
of the substitution of a hydroxy group in the acyl
moiety of capsaicin has not been reported. Only one
case has demonstrated that the efficacy of TRPV1
activity was increased without greatly altering potency
when a hydroxyl group was substituted in the longer
acyl moiety (C18:1) of vanilloid.12) The vanilloid-bind-
ing site of TRPV1 has been determined by means of
the gene engineering technique.30,31) The recognition
site of the vanillyl moiety in vanilloids is located in
the N- and C-terminal cytosolic tails of the TRPV1
membrane protein. On the other hand, the acyl moiety-
binding site located in the transmembrane region of
TRPV1 requires the acyl moiety to have a specific
hydrophobicity. Therefore, the decrease in hydropho-
bicity due to the hydroxylation of the acyl moiety in
capsaicin (viz., capsaicinol) probably reduces its affinity
to TRPV1. Furthermore, recognition of the side chain of
vanilloids by TRPV1 might be loose because there are
various TRPV1 activators that possess diverse side
chains. This loose recognition might explain the equal
effectiveness of TRPV1 activation by the enantiomers
of capsaicinol.
Figure 5A shows the time course of adrenaline
secretion in anesthetized rats intravenously injected
with (ꢁ)-capsaicinol. A dose of over 2.5 mg/kg (8.0
mmol/kg body weight) (ꢁ)-capsaicinol enhanced ad-
renaline secretion immediately after injection. Once
secretion declined to half the initial response, it gradual-
ly increased again. In brief, (ꢁ)-capsaicinol showed a
two-phase pattern of adrenaline secretion, i.e., rapid and
transient increases and slow and prolonged increases.
These responses increased dose-dependently. The initial
response elicited by a dose of 5.0 mg/kg (ꢁ)-capsaicinol
was comparable to that of a 0.05 mg/kg dose of
capsaicin. On the other hand, the later response even
when a dose of 2.5 mg/kg (ꢁ)-capsaicinol was injected
tended to be greater than when a dose of 0.05 mg/kg
capsaicin was injected. The total amount of adrenaline
secretion over 18 min caused by 0 (vehicle), 0.5, 2.5, and
5.0 mg/kg (ꢁ)-capsaicinol, and a 0.05 mg/kg dose of
capsaicin, was 240, 270, 1,150, 1,640, and 1,190 ng/kg
respectively. Coinjection of a dose of 20 mg/kg cap-
sazepine significantly attenuated the response to a 2.5
mg/kg dose of (ꢁ)-capsaicinol (Fig. 5B).
Previously, we found that some vanilloids such as
capsaicin, olvanil, and resiniferatoxin caused adrenaline
secretion in rats.18) There were two time-dependent
patterns of secretion: a capsaicin-like two-phase pattern
and an olvanil-like one-phase pattern. The response to
capsaicinol showed the former pattern. This is probably

Synthesis and TRPV1-Related Actions of Capsaicinol
1911
because capsaicinol is structurally similar to capsaicin
rather than to olvanil. Kawada et al. reported that the
biological half-life of intraperitoneally injected capsai-
cin in blood was approximately 7 min in rats.32) Intra-
venously injected capsaicin was reduced by half at 4 min
(Watanabe et al., unpublished data). Therefore, the first
phase of adrenaline secretion, up to 9 min caused by
capsaicin and capsaicinol probably depends on their
concentrations in the blood. The maximum secretion
observed immediately after injection of capsaicin was
saturated at a dose of 0.1 mg/kg.18) The maximum
secretion induced by a dose of 5 mg/kg capsaicinol was
equivalent to that induced by 0.05 mg/kg of capsaicin,
which was half of the saturated secretion induced by
capsaicin. On the other hand, the total amount of
adrenaline secretion induced by 0.05 mg/kg of capsaicin
was comparable to that induced by 2.5 mg/kg of
capsaicinol. Therefore, the effectiveness of capsaicinol
in inducing adrenaline secretion was approximately 50–
100 times lower than that of capsaicin. The relationship
between the potencies of capsaicinol and capsaicin
coincided with that of the above-mentioned TRPV1
activation potencies. The agreement of physiological
responses in the whole body and at the molecular level is
probably due to similar physicochemical properties
based on the structural resemblance between the two
compounds.
In other words, the potency of a ligand depends not only
on its affinity but also on the accessibility to its receptor.
Thus the physicochemical properties of ligands, such as
lipophilicity and stability, are important in understand-
ing the response. Extremely low or high lipophilicity
reduces pungency even if the ligand possesses a high
affinity for TRPV1. For example, olvanil, a capsaicin
analog having a longer acyl chain, possesses greater
TRPV1 activation potency than capsaicin, in spite of its
non-pungency.13) It is believed that the extremely high
lipophilicity (log P ¼ 9:12) of olvanil hardly allows it
access to TRPV1 in the tongue because TRPV1 is
covered with the epitheliocyte in the tongue.33) Capsiate,
which has a higher lipophilicity (log P ¼ 5:80) than
capsaicin, is also non-pungent.13) In case of capsiate, its
lipophilicity and instability probably contribute to its
organoleptic properties.34) In contrast to olvanil and
capsiate, the lipophilicity of capsaicinol is lower than
that of capsaicin, but the accessibility of capsaicinol to
TRPV1 in the tongue might not differ greatly from that
of capsaicin because capsaicinol can still be considered
lipophilic (log P > 1). Considering that the accessibility
and stability of capsaicinol are identical to those of
capsaicin, it is not surprising that the relative pungency
of capsaicinol to capsaicin is consistent with the relative
TRPV1 activation potency of capsaicinol to capsaicin.
The same phenomenon probably explains the relation-
ship between TRPV1 activation potency and the adrena-
line secretion induced by the two compounds.
The relationship between the physiological activities
and lipophilicity of capsaicinol
We have described the effects of capsaicinol on
TRPV1 and adrenaline secretion in the present paper.
Additionally, we evaluated the pungency of capsaicinol
as one of the TRPV1-related responses. Table 1 shows
the pungency thresholds of capsaicinol and capsaicin
evaluated by two methods. The results from both
methods indicated that the pungency of capsaicinol
was approximately 100-fold weaker than that of cap-
saicin. On the other hand, we measured the lipophilicity,
log P, of the two compounds as one of the physico-
chemical characters by the n-octanol/water flask shak-
ing method.24) As Table 1 shows, the lipophilicity of
capsaicinol is lower than that of capsaicin, which is in
agreement with our expectations.
References
1) Nakatani, N., Tachibana, Y., and Kikuzaki, H., Anti-
oxidative compounds from edible plants: phenolic
amides from Capsicum frutescens L. In ‘‘Medical,
Biochemical and Chemical Aspects of Free Radicals,’’
Elsevier Science Publishers, Amsterdam, p. 453 (1989).
2) Masuda, T., Kikuzaki, H., Tachibana, Y., and Nakatani,
N., Synthesis of capsaicinol [1]. Chem. Express, 5, 369–
372 (1990).
3) Masuda, T., and Nakatani, N., Synthesis and absolute
structure of (ꢃ)-capsaicinol. Agric. Biol. Chem., 55,
2337–2340 (1991).
4) Buck, S. H., and Burks, T. F., The neuropharmacology
of capsaicin: review of some recent observations.
Pharmacol. Rev., 38, 179–226 (1986).
5) Govindarajan, V. S., and Sathyanarayana, M. N.,
Capsicum-production, technology, chemistry, and qual-
ity. V. Impact on physiology, pharmacology, nutrition,
and metabolism: structure, pungency, pain, and desensi-
tization sequences. CRC Crit. Rev. Food Sci. Nutr., 29,
435–474 (1991).
6) Caterina, M. J., Schumacher, M. A., Tominaga, M.,
Rosen, T. A., Levine, J. D., and Julius, D., The capsaicin
receptor: a heat-activated ion channel in the pain
pathway. Nature, 389, 816–824 (1997).
To show physiological activity, a ligand must gain
access to its receptor across certain ‘‘barriers’’ such as
the mucosa, lyase enzyme, its own stability, and so on.
Table 1. Pungency (Oil Dilution Values [ODV] and Scoville Heat
Units [SHU]) and Lipophilicity (log P) of Capsaicin and (ꢁ)-
Capsaicinol
Pungency
log P
ODV (ꢂ104)
SHU (ꢂ106)
7) Clapham, D. E., TRP channels as cellular sensors.
Nature, 426, 517–524 (2003).
8) Szallasi, A., and Blumberg, P. M., Vanilloid (capsaicin)
receptors and mechanisms. Pharmacol. Rev., 51, 159–
(ꢁ)-Capsaicinol
0.18
20
0.12
16ꢄ
1.74
3.81
Capsaicin
ꢄThe SHU of capsaicin is quoted from the literature.35)

1912
K. KOBATA et al.
211 (1999).
21) Iwai, K., Yazawa, A., and Watanabe, T., Roles as
metabolic regulators of the non-nutrients, capsaicin and
capsiate, supplemented to diets. Proc. Jpn. Acad., 79B,
207–212 (2003).
22) Sugai, E., Morimitsu, Y., Iwasaki, Y., Morita, A.,
Watanabe, T., and Kubota, K., Pungent qualities of
sanshool-related compounds evaluated by a sensory test
and the activation of rat TRPV1. Biosci. Biotechnol.
Biochem., 69, 1951–1957 (2005).
23) ISO 3513, Spices and condiments—Chillies—Determi-
nation of Scoville index (1977).
24) OECD Partition coefficient (n-octanol/water)-flaskshak-
ing method, OECD guideline for testing chemicals 107
(1981).
25) Harris, T. M., and Carney, R. L., Allylic oxidation of
olefins by palladium acetate. J. Am. Chem. Soc., 88,
2054–2055 (1966).
26) Rappoport, Z., Winstein, S., and Young, W. G., Allylic
oxidation of olefins by mercuric acetate. J. Am. Chem.
Soc., 94, 2320–2329 (1972).
27) Gilmore, J. R., and Mellor, J. M., Oxidation of olefins by
manganese(III) acetate to give allylic acetates. J. Chem.
Soc. C: Organic, 1971, 2355–2357 (1971).
28) Janusz, J. M., Buckwalter, B. L., Young, P. A., LaHann,
T. R., Farmer, R. W., Kasting, G. B., Loomans, M. E.,
Kerckaert, G. A., Maddin, C. S., Berman, E. F., Bohne,
R. L., Cupps, T. L., and Milstein, J. R., Vanilloids. 1.
Analogs of capsaicin with antinociceptive and anti-
inflammatory activity. J. Med. Chem., 36, 2595–2604
(1993).
9) Witte, D. G., Cassar, S. C., Masters, J. N., Esbenshade,
T., and Hancock, A. A., Use of a fluorescent imaging
plate reader-based calcium assay to assess pharmaco-
logical differences between the human and rat vanilloid
receptor. J. Biomol. Screen., 7, 466–475 (2002).
10) Yang, B. H., Piao, Z. G., Kim, Y. B., Lee, C. H., Park,
K., Kim, J. S., and Oh, S. B., Activation of vanilloid
receptor 1 (VR1) by eugenol. J. Dent. Res., 82, 781–785
(2003).
11) Melck, D., Bisogno, T., De Petrocellis, L., Chuang, H.,
Julius, D., Bifulco, M., and Di Marzo, V., Unsaturated
long-chain N-acyl-vanillyl-amides (N-AVAMs): vanil-
loid receptor ligands that inhibit anandamide-facilitated
transport and bind to CB1 cannabinoid receptors.
Biochem. Biophys. Res. Commun., 262, 275–284 (1999).
12) Appendino, G., and Minassi, A., N-Acylvanillamides:
development of an expeditious synthesis and discovery
of new acyl templates for powerful activation of the
vanilloid receptor. J. Med. Chem., 45, 3739–3745
(2002).
13) Iida, T., Moriyama, T., Kobata, K., Morita, A.,
Murayama, N., Hashizume, S., Fushiki, T., Yazawa, S.,
Watanabe, T., and Tominaga, M., TRPV1 activation and
induction of nociceptive response by a non-pungent
capsaicin-like compound, capsiate. Neuropharmacology,
44, 958–967 (2003).
14) Watanabe, T., Kawada, T., Yamamoto, M., and Iwai, K.,
Capsaicin, a pungent principle of hot red pepper, evokes
catecholamine secretion from the adrenal medulla of
anesthetized rats. Biochem. Biophys. Res. Commun., 142,
259–264 (1987).
29) Klopman, G., and Li, J. Y., Quantitative structure-
agonist activity relationship of capsaicin analogues. J.
Comput. Aids. Mol. Des., 9, 283–294 (1995).
30) Jordt, S. E., and Julius, D., Molecular basis for species-
specific sensitivity to ‘‘hot’’ chili peppers. Cell, 108,
421–431 (2002).
15) Kobayashi, A., Osaka, T., Namba, Y., Inoue, S., Lee, T.
H., and Kimura, S., Capsaicin activates heat loss and
heat production simultaneously and independently in
rats. Am. J. Physiol., 44, R92–R98 (1998).
16) Kawada, T., Hagihara, K., and Iwai, K., Effects of
capsaicin on lipid metabolism in rats fed a high fat diet.
J. Nutr., 116, 1272–1278 (1986).
17) Watanabe, T., Kawada, T., Kato, T., Harada, T., and
Iwai, K., Effects of capsaicin analogs on adrenal
catecholamine secretion in rats. Life Sci., 54, 369–374
(1994).
18) Watanabe, T., Sakurada, N., and Kobata, K., Capsaicin-,
resiniferatoxin-, and olvanil-induced adrenaline secre-
tions in rats via the vanilloid receptor. Biosci. Biotech-
nol. Biochem., 65, 2443–2447 (2001).
19) Kim, K. M., Kawada, T., Ishihara, K., Inoue, K., and
Fushiki, T., Swimming capacity of mice is increased by
oral administration of a nonpungent capsaicin analog,
stearoyl vanillylamide. J. Nutr., 128, 1978–1983 (1998).
20) Ohnuki, K., Haramizu, S., Oki, K., Watanabe, T.,
Yazawa, S., and Fushiki, T., Administration of capsiate,
a non-pungent capsaicin analog, promotes energy me-
tabolism and suppresses body fat accumulation in mice.
Biosci. Biotechnol. Biochem., 65, 2735–2740 (2001).
31) Jung, J., Lee, S. Y., Hwang, S. W., Cho, H., Shin, J.,
Kang, Y. S., Kim, S., and Oh, U., Agonist recognition
sites in the cytosolic tails of vanilloid receptor 1. J. Biol.
Chem., 277, 44448–44454 (2002).
32) Kawada, T., Watanabe, T., Katsura, K., Takami, H., and
Iwai, K., Formation and metabolism of pungent principle
of Capsicum fruits xv. J. Chromatogr., 329, 99–105
(1985).
33) Ishida, Y., Ugawa, S., Ueda, T., Murakami, S., and
Shimada, S., Vanilloid receptor subtype-1 (VR1) is
specifically localized to taste papillae. Mol. Brain Res.,
107, 17–22 (2002).
34) Sutoh, K., Kobata, K., and Watanabe, T., Stability of
capsinoid in various solvents. J. Agric. Food Chem., 49,
4026–4030 (2001).
35) Todd, P. H., Jr., Bensinger, M. G., and Biftu, T.,
Determination of pungency due to capsicum by gas-
liquid chromatography. J. Food Sci., 42, 660–665
(1977).