Structural analogues of homoeriodictyol as flavor modifiers. Part III: Short chain gingerdione derivatives

Article abstract of DOI:10.1021/jf8006536

In order to find new flavor modifiers, various short chain gingerdione derivatives were synthesized as structural analogues of the known bitter masker homoeriodictyol and evaluated by a sensory panel for masking and sweetness enhancing activities. 1-(4-Hy

Full text of DOI:10.1021/jf8006536

6656 J. Agric. Food Chem. 2008, 56, 6656–6664
Structural Analogues of Homoeriodictyol as Flavor
Modifiers. Part III: Short Chain Gingerdione
Derivatives
¨
JAKOB P. LEY,* SUSANNE PAETZ, MARIA BLINGS, PETRA HOFFMANN-LUCKE,
¨
HEINZ-JURGEN BERTRAM, AND GERHARD E. KRAMMER
Symrise GmbH & Co. KG, Flavor & Nutrition Research & Innovation, P.O. Box 1253,
37601 Holzminden, Germany
In order to find new flavor modifiers, various short chain gingerdione derivatives were synthesized
as structural analogues of the known bitter masker homoeriodictyol and evaluated by a sensory panel
for masking and sweetness enhancing activities. 1-(4-Hydroxy-3-methoxyphenyl)hexa-3,5-dione ([2]-
gingerdione) and the homologue 1-(4-hydroxy-3-methoxyphenyl)hepta-3,5-dione ([3]-gingerdione) at
concentration ranges 50-500 mg kg-1 showed the most promising masking activity of 20-30%
against bitterness of a 500 mg kg-1 aqueous caffeine solution. Additionally, both compounds were
able to reduce the bitterness of a 5 mg kg-1 quinine solution by about 20%; however, the bitter
tastes of salicine, the model peptide H-Leu-Trp-OH, and KCl solutions were not reduced. Whereas
for bitter masking activity a vanillyl moiety seems to be important, some of the tested isovanillyl isomers
showed an interesting sweet enhancing effect without exhibiting a significant intrinsic sweetness.
The isomer 1-(3-hydroxy-4-methoxyphenyl)hexa-3,5-dione ([2]-isogingerdione) at 100 mg kg-1 caused
a significant and synergistic increase of 27% of sweet taste of a 5% sucrose solution.
KEYWORDS: Taste modifying; bitter masking; sweet enhancing; gingerdiones
INTRODUCTION
mostly achieved by using high impact sweeteners such as
saccharin, aspartam, or acesulfam and/or bulk calorie free
sweeteners such as sorbitol or erythritol (12). But very recently,
the use of sweet enahncers such as pyridinium alanyl betaine
(13) or ethyl butyrate in subthreshold concentrations (14) was
demonstrated. Both molecules can increase the perceived
sweetness of a sugar containing application without exhibiting
strong intrinsic sweetness.
Recently, we identified some hydroxybenzoic acid vanillyl-
amides (15) and hydroxylated deoxybenzoins (16) as new bitter
masking compounds related to the active flavanone homoeri-
odictyol (1), which was identified as the active principle of the
long known bitter reducing extract from Herba Santa (17). Some
of the new vanillylamides such as 2,4-dihydroxybenzoic acid
N-vanillylamide (2) showed additionally sweet enhancing
properties (18) and may be used in flavors to increase preference
of low-carbohydrate applications.
Starting with the results from former structure-activity
relationships (17, 15) we tried to simplify the homoeriodictyol
structure and have investigated into the promising class of short
chain gingerdiones and related molecules (Figure 1) to discover
new taste modulators. Longer chain gingerdiones (1-(4-hydroxy-
3-methoxyphenyl)-3,5-alkadiones) are minor constituents of the
ginger rhizomes in addition to paradols (1-(4-hydroxy-3-
methoxyphenyl)-3-alkanones), gingerols (1-(4-hydroxy-3-meth-
oxyphenyl)-5-hydroxy-3-alkanones), and shogaols (1-(4-hydroxy-
3-methoxyphenyl)-5E-alken-3-ones) (19). In dry ginger, [6]-,
Masking bitter and other negative taste attributes as well as
enhancing positive taste qualities such as sweetness is crucial
for the food and beverage industry especially in improving the
palatability of functional foods enriched with bitter tasting
actives such as tea catechins (1) or reduced in calories, especially
in sucrose or high fructose corn syrup. Several possibilities to
counteract bitter taste are known: for example, removal of bitter
actives, physical barriers (e.g., (micro)encapsulation, coatings,
emulsions, suspensions) (2), using lipoproteins (3), proteins (4),
polysaccharides such as pectins (5) or cyclodextrins (6) for
scavenging the tastants, biotransformation of bitter molecules
to nonbitter metabolites (e.g., naringin hydrolysis to naringenin
(7)), using strong flavors or tastants (e.g., salt, sweeteners, acids)
or congruent flavors (e.g., chocolate flavor, grapefruit flavor etc.)
(8), and last but not least, bitter masking molecules, which are
somewhat odorless and tasteless (9). In recent years some potent
new bitter masking molecules were identified: AMP shows a
well accepted masking effect against KCl bitterness in sodium
reduced formulations (10), and a pyridinium glycinyl betaine
was able to reduce the bitterness ratings of various concentra-
tions of caffeine (250-2500 mg kg-1) by about 3 units using
a scale of 0 (no bitterness) to 5 (very strong) (11). Until now,
maintaining the sweet taste of calorie reduced applications was
* To whom correspondence should be addressed. Phone: + 49 5531
90 1883. Fax: + 49 5531 90 48883. E-mail: jakob.ley@symrise.com.
10.1021/jf8006536 CCC: $40.75  2008 American Chemical Society
Published on Web 07/04/2008

Short Chain Gingerdione Derivatives as Flavor Modifiers
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6657
7.51* (ca. 5% H, d, J ) 16.1 Hz, H-1*), 7,11* (ca. 5% H, ddd, J )
8.2 Hz, J ) 2 Hz, J ) 0.5 Hz), 7.09 (1H, ddd, J ) 8.2 Hz, J ) 1.9 Hz,
J ) 0.5 Hz, H-6′), 7.05* (5% H, d, J ) 2 Hz), 7.02 (1H, d, J ) 1.9
Hz, H-2′), 6.93* (5% H, d, J ) 8.2 Hz), 6.92 (1H, d, J ) 8.2 Hz,
H-5′), 6.63* (ca. 5% H, d, J ) 16.1 Hz), 6.32 (1H, d, J ) 15.8 Hz,
H-2), 5.91 (1H, bs, OH), 5.62 (1H, s, H-4), 3.94 (3H, s, O-CH3), 2.16
(3H, s, H-6) ppm. 13C NMR (100 MHz, CDCl3 internal standard TMS):
δ ) 197.03 (C, C-5), 177.93 (C, C-3), 147.72 (C, C-3′ or C-4′), 146.79
(C, C-4′ or C-3′), 140.06 (CH, C-1), 127.67 (CH, C-1′), 122.65 (CH,
C-6′), 120.32 (CH, C-2), 114.82 (CH, C-5′), 109.50 (CH, C-2′), 100.71
(CH, C-4), 55.95 (CH3, O-CH3), 26.58 (CH3, C-6) ppm.
1-(4-Hydroxy-3-methoxyphenyl)hexa-3,5-dione (6). (1E)-1-(4-Hy-
droxy-3-methoxyphenyl)hex-1-ene-3,5-dione (5) (2.28 g) was mixed
with palladium on activated carbon (10%, wet, 0.23 g) in ethanol (100
mL) and was hydrogenated at ambient pressure and room temperature
with hydrogen for 2 h (410 mL H2 uptake). The reaction mixture was
filtered, and the filtrate was evaporated in vacuo. The raw material was
purified using FC on silica gel 60 using the eluent n-hexane/ethyl acetate
2:1 (v/v). Yield: 1.16 g (50%) of colorless oil (slowly crystallizing
over months). HPLC-MS (RP-18-Phase, ESI+): m/z ) 236.92 (100%,
[M + H]+), 219.08 (94%, [M - HO]+). HRMS: calcd. for C13H16O4
236.1049, found 236.1041. 1H NMR (400 MHz, CDCl3, internal
standard TMS, *diketo tautomer): δ ) 6.83 (1H, dd, J ) 7.2 Hz, J )
0.8 Hz, H-5′), 6.82* (ca. 10% H, m, H-5′), 6.69 (1H, m, H-2′), 6.67
(1H, ddm, J ) 7.2 Hz, J ) 1.6 Hz, H-6′), 5.53 (1H, s, OH), 5.47 (1H,
s, H-4), 3.88* (10% H, s, O-CH3), 3.86 (3H, s, O-CH3), 3.54* (10%
2H, s, H-4*), 2.86 (2H, dd, J ) 8.4 Hz, J ) 7.2 Hz), 2.82* (10% 2H,
ddd, J ) 6.8 Hz, J ) 4.4 Hz, J ca. 1 Hz), 2.559* (10% 2H, m), 2.558
(2H, dd, J ) 8.4 Hz, 8.4 Hz), 2.20* (10% 3H, s, H-6*), 2.03 (3H, s,
H-6) ppm. 13C NMR (100 MHz, CD3OD, internal standard TMS): δ
) 203.38* (C, C-3*), 193.27 (C, C-3), 191.18 (C, C-5), 146.42 (C,
C-3′), 144.00 (C-4′), 132.62 (C-1′), 120.82 (C-6′), 114.38* (CH2, C-5′*),
114.34 (CH2, C-5′), 111.01* (CH2, C-2′*), 110.96 (CH2, C-2′), 100.13
(CH2, C-4), 58.13* (CH3, O-CH3*), 55.88 (CH3, O-CH3), 45.53* (CH2,
C-4*), 40.40 (CH2, C-2), 31.29 (CH2, C-1), 29.15* (CH2), 24.86 (CH3,
C-6) ppm.
(1E)-1-(3-Hydroxy-4-methoxyphenyl)hex-1-ene-3,5-dione (7). The
dehydrogingerdione 7 was synthesized analogously to 5 using boron
trioxide (7.0 g), acetyl acetone (14.3 g), tributyl borate (42.6 g),
isovanillin (7.05 g). and butylamine (2 mL). The raw material was
recrystallized from ethyl acetate. Yield: 4.95 g of yellow-orange crystals
(purity >95%, LC-MS). For analytics, a small sample was purified
with FC on silica gel 60 using the eluent n-hexane/ethyl acetate 1:1
(v/v) (99.7%, GC). HPLC-MS (RP-18-Phase, ESI+): m/z ) 235.0
(100%, [M + H]+), 177.4 (18%). HRMS: calcd. for C13H14O4 234.0892,
found 234.0888. 1H NMR (400 MHz, CDCl3, internal standard TMS):
δ ) 7.51 (1H, d, J ) 15.79 Hz, H-1), 7.14 (1H, d, J ) 2.1 Hz, H-2′),
7.02 (1H, ddd, J ) 8.3 Hz, J ) 2.1 Hz, J ) 0,4 Hz, H-6′), 6.85 (1H,
d, J ) 8.3 Hz, H-5′), 5.69 (1H, s, OH), 6.32 (1H, d, J ) 15.8 Hz,
H-2), 5.62 (1H, s, H-3), 3.92 (3H, s, O-CH3), 2.16 (3H, s, H-6) ppm.
13C NMR (100 MHz, CDCl3, internal standard TMS): δ ) 197.30 (C,
C-5), 177.67 (C, C-3), 148.31 (C, C-4′), 145.91 (C, C-3′), 139.72 (CH,
C-1), 128.76 (CH, C-1′), 121.86 (CH, C-6′), 120.92 (CH, C-2), 112.67
(CH, C-5′), 110.89 (CH, C-2′), 100.89 (CH, C-4), 56.02 (CH3, O-CH3),
26.91 (CH3, C-6) ppm.
Figure 1. Possible simplifications and modifications of the homoeriodictyol
structure.
[8]-, [10]-, and [12]-gingerdiones (20) and in fresh ginger [6]-,
[8]-, and [10]-dehydrogingerdiones (1-(4-hydroxy-3-methoxy-
phenyl)-1-alkene-3,5-diones) were described (21). Only one
short chain derivative, the [3]-dehydrogingerdione (11, Figure
2), was found in ginger as a trace component (21). The
antioxidant hispolon 1-(3,4-dihydroxyphenyl)-1E-hex-1-en-3,5-
dione (13, Figure 2) was identified in the fungus Inonotus
hispidus (22), but it was never tested for taste effects. In our
studies, we have focused on the short chain derivatives due to
the intrinsic pungency of most of the longer chain ginger
vanilloids. Several new and known relatives were synthesized
and screened for their masking effects against caffeine and
enhancing effects of sucrose sweetness.
MATERIALS AND METHODS
Dipeptide N-L-leucyl-L-tryptophane (H-Leu-Trp-OH) was obtained
from Bachem (Weil am Rhein, Germany). Homoeriodictyol (1) was
isolated as a sodium salt as described in ref (17), 3 was from ABCR
(Karlsruhe, Germany, order no. TCH0996), 4 was from the Symrise
library (prod. no. 113507), curcumin (19) was purchased from Sigma-
Aldrich (Deisenhofen, Germany), and stevioside from Jenaer Pflan-
zenprodukte (Jena, Germany). [8]-Gingerol (15) and [10]-gingerol (16)
were isolated in a purity >95% from a commercial ginger oleoresin
(Naturex, Avignon, France) using common PHPLC techniques. All
other chemicals were from Sigma-Aldrich, Acros (Schwerte, Germany),
E. Merck (Darmstadt, Germany), or Lancaster Synthesis (Frankfurt,
Germany). NMR spectra were recorded using a Varian VXR400S (1H:
400 MHz) spectrometer (Varian, Darmstadt, Germany) at 25 °C using
tetramethyl silane as the internal standard. LC-MS spectra were recorded
using the LCQ HPLC system Finnigan MAT HP1100 (Finnigan MAT,
Egelsbach, Germany; APCI, atmospheric pressure chemical ionization).
Elemental analysis (combustion analysis for C, H, and O and AAS for
sodium) was performed by Bayer Technology Services, Germany. Flash
chromatography (FC) was performed on a Biotage Flash 40 system
(Biotage AB, Uppsala, Sweden) using disposable prep-packed columns.
Solvents were dried as needed by using an activated molecular
sieve.
(1E)-1-(4-Hydroxy-3-methoxyphenyl)hex-1-ene-3,5-dione (5). Af-
ter stirring boron trioxide (3.76 g) and acetyl acetone (7.58 g) at 100
°C for 1 h, the mixture was cooled to 0 °C. Tributyl borate (23.0 g)
was dissolved in dry ethyl acetate (50 mL) and added. At 0 °C, first
vanillin (3.76 g) dissolved in dry ethyl acetate (30 mL) and then
butylamine (0.5 mL) in dry ethyl acetate (3 mL) were added dropwise.
The mixture was slowly warmed to room temperature and stirred for
3 days. Hydrochloric acid (0.4 mol/L, 50 mL, 60 °C) was added, and
the phases separated. The organic phase was washed with water, dried
with Na2SO4, filtered, and evaporated in vacuo. The raw material (6.6
g) was purified by flash chromatography on silica gel 60 using the
eluent n-hexane/ethyl acetate 1:1 (v/v). Yield: 2.97 g (51% related to
vanillin) of yellow oil. HPLC-MS (RP-18-Phase, APCI +): m/z )
235.19 (100%, [M + H]+). 1H NMR (400 MHz, CDCl3, internal
standard TMS, *diketo tautomer): δ ) 7.53 (1H, d, J ) 15.8 Hz, H-1),
1-(3-Hydroxy-4-methoxyphenyl)hexa-3,5-dione (8). The hydro-
genation of 7 (2.34 g) to 8 was performed as described for [2]-ging-
erdione (6) (g). The crude reaction product (69%, GC) was purified
with FC on silica gel 60 using the eluent n-hexane/ethyl acetate 2:1
(v/v). Yield: 1.0 g (42%, purity 99.0% GC) of colorless crystals. MS
(EI): m/z ) 236 (M+ ·, 43%), 150 (14%), 137 (100%), 85 (15%), 43
(17%), 28 (14%). HRMS: calcd. for C13H16O4 236.1049, found
236.1044. 1H NMR (400 MHz, CDCl3, internal standard TMS, *diketo
tautomer): δ ) 6.76 (1H, d, J ) 8.2 Hz, H-5′), 6.77 (1H, d, J ) 2.1
Hz, H-2′), 6.62 (1H, ddm, J ) 8.2 Hz, J ) 2.1 Hz, H-6′), 5.60 (1H, s,
OH), 5.47 (1H, s, H-4), 3.861 (3H, s, O-CH3), 3.856* (ca. 5% H, s,
O-CH3*), 2.85 (2H, m, H-1), 2.55 (2H, dd, J ) 9 Hz, 9 Hz), 2.20* (ca.
5% 3H, H-6*, s), 2.04 (3H, s, H-6) ppm. 13C NMR (100 MHz, CD3OD,
internal standard TMS): δ ) 193.30 (C, C-3), 191.09 (C, C-5), 145.56
(C, C-3′), 145.03 (C, C-4′), 134.00 (C, C-1′), 119.68* (CH, C-6′*),
119.65 (CH, C-6′), 114.38* (CH2, C-5′*), 114.34 (CH2, C-5′), 110.75*

6658 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Ley et al.
Figure 2. Synthesized and tested compounds related to gingerdiones.
(CH, C-2′*), 110.70 (CH, C-2′), 100.13 (CH, C-4), 58.7* (CH3,
O-CH3*), 56.00 (CH3, O-CH3), 45.34* (CH2, C-4*), 40.10 (CH2, C-2),
30.89 (CH2, C-1), 28.84* (CH2), 24.85 (CH3, C-6) ppm.
°C for 1 h, 55 mL of the solvent was evaporated in vacuo, and water
(250 mL) was added. The mixture was stirred at 60 °C for another
hour and at room temperature for 12 h. The reaction mixture was
extracted with ethyl acetate (200 mL), and the solvents were evaporated.
The crude reaction product was suspended in ethyl acetate (50 mL)
and stored for several days. The filtered solution (the precipitate contains
mainly the curcumin derivative) was purified by FC on silica gel 60
using the eluent ethyl acetate/hexane (25 min 30:70 (v/v), then 10 min
isocratic to 50:50). The isolated dehydrogingerdione-[3] (11) was
recrystallized from ethyl acetate. Yield: 7 g (67%) of yellow crystals.
MS (EI): m/z ) 248 (M+ ·, 55%), 230 (28%), 219 (49%), 201 (50%),
191 (63%), 177 (100%), 159 (15%), 145 (50%). HRMS: calcd. for
C16H14O4 248.1049, found 248.1028. 1H NMR (400 MHz, CDCl3,
internal standard TMS): δ ) 7.52 (1H, d, J ) 15.7 Hz, H-1), 7.09
(1H, ddd, J ) 8.2 Hz, J ) 1.9 Hz, J ) 0.5 Hz, H-6′), 7.02 (1H, d, J
) 1.9 Hz, H-2′), 6.92 (1H, d, J ) 8.2 Hz, H-5′), 6.34 (1H, d, J ) 15.8
Hz, H-2), 5.86 (1H, bs, 4′-OH), 5.63 (1H, s, 3-OH), 3.94 (3H, s, OCH3),
(1E)-1-(4-Hydroxyphenyl)hex-1-ene-3,5-dione (9). The dehydro-
gingerdione 9 was synthesized analogously to 5 using boron trioxide
(13.9 g), acetyl acetone (28.4 g), tributyl borate (85.2 g), 4-hydroxy-
benzaldehyde (12.3 g), and butylamine (2 mL). The raw material was
recrystallized from ethyl acetate. Yield: 2.5 g of yellow-orange crystals
(purity >95%, LC-MS). MS (EI): m/z ) 204 (100%, M+·), 186 (40%),
161 (75%), 147 (100%), 119 (32%), 91 (28%, 65 (20%), 43 (43%).
HRMS: calcd. for C12H12O3 248.1049, found 248.1028. 1H NMR (400
MHz, CDCl3, internal standard TMS, *diketo tautomer): δ ) 7.53 (2H,
m, H-2′, H-6′), 7.49 (1H, d, J ) 16 Hz, H-1), 6.81 (2H, m, H-3′, H-5′),
6.56* (13% H, d, J ) 16.1 Hz, H-2*), 6.58 (1H, d, J ) 16.0 Hz, H-2),
5.85 (1H, s, H-4), 2.18* (13% 3H, s, H-6*), 2.12 (3H, s, H-6) ppm.
13C NMR (100 MHz, CD3OD, internal standard TMS): δ ) 196.53
(C, C-5), 178.19 (C, C-3), 159.56 (C, C-4′), 139.82 (CH, C-1), 130.57*
(CH), 130.02 (2 × CH, C-2′, C-6′), 125.71 (C, C-1′), 119.31 (CH,
C-2), 115.82* (CH), 115.76 (2 × CH, C-3′, C-5′), 100.39 (CH, C4),
26.29 (CH3, C-6) ppm.
1-(4-Hydroxyphenyl)hexa-3,5-dione (10). The hydrogenation of 9
(2.00 g) to 10 was performed as described for [2]-gingerdione (6). The
reaction product was purified with FC on silica gel 60 using the eluent
n-hexane/ethyl acetate 2:1 (v/v). Yield: 0.5 g (25%) of colorless oil.
MS (EI): m/z ) 206 (M+ ·, 32%), 120 (23%), 107 (100%), 85 (21%),
77 (12%), 43 (17%). HRMS: calcd. for C12H14O3 206.0943, found
206.0933. 1H NMR (400 MHz, CDCl3, internal standard TMS, *diketo
tautomer): δ ) 7.02 (2H, d, J ) 8.0 Hz, H-2′, 6′), 7.00* (20% 2H, d,
8.0 Hz, H-2′*, 6′*), 6.76 (2H, d, J ) 8.0 Hz, H-3′, 5′), 6.74* (2H, m,
H-3′*, 5′*), 5.48 (1H, s, H-4), 3.56* (20% 2H, s, H-4*), 2.85 (2H, m,
H-1), 2.80* (2H, m, J ) 7.1 Hz, J ) 4.5 Hz, H-1*), 2.55 (2H, dd, J )
8.4 Hz, J ) 7.2 Hz, H-2), 2.19* (1H, s, H-6*), 2.04 (3H, s, H-6), 1.1*
(20% H, 3-OH*) ppm. 13C NMR (100 MHz, CD3OD, internal standard
TMS): δ ) 204.15* (C, C-3*), 203.08* (C, C-5*), 193.37 (C, C-3),
191.96 (C, C-5), 154.37* (C, C-4′*), 154.33 (C, C-4′), 132.32 (C, C-1′),
132.08* (C, C-1′*), 129.39* (2 × CH, C-2′*, 6′*), 129.36 (2 × CH,
C-2′, 6′), 115.48* (2 × CH, C-3′*, 5′*), 115.41 (2 × CH, C-3′, 5′),
100.21 (CH, C-4), 57.98* (CH2, C-4*), 45.51* (CH2, C-2*), 40.20 (CH2,
C-2), 30.94* (CH3, C-6*), 30.78 (CH2, C-1), 28.58* (CH2, C-1*), 24.97
(CH3, C-6) ppm.
2.42 (2H, q, J ) 7.5 Hz, H-6), 1.18 (3H, t, J ) 7.5 Hz, H-7) ppm. 13
C
NMR (100 MHz, CD3OD, internal standard TMS): δ ) 201.00 (C,
C-5), 177.70 (C, C-3), 147.66 (C, C-4′), 146.78 (C, C-3′), 139.77 (CH,
C-1), 127.75 (C, C-1′), 122.61 (CH, C-6′), 120.50 (CH, C-2), 114.81
(CH, C-5′), 109.47 (CH, C-2′), 99.55 (CH, C-2), 55.96 (CH3, O-CH3),
33.23 (CH2, C-6), 9.45 (CH3, C-7) ppm.
1-(4-Hydroxy-3-methoxyphenyl)hepta-3,5-dione (12). The hydro-
genation of 11 (3.5 g) to 12 was performed as described for
[2]-gingerdione (6). The reaction product was purified with FC on silica
gel 60 using the eluent n-hexane/ethyl acetate 2:1 (v/v). Yield: 2.9 g
(83%) of colorless oil (GC DB1 RI 2002 96.6%). MS (EI): m/z ) 250
(M+ ·, 30%), 150 (12%), 137 (100%), 99 (10%), 47 (8%). HRMS:
1
calcd. for C16H14O4 248.1049, found 248.1028. H NMR (400 MHz,
CDCl3, internal standard TMS, *diketo tautomer): δ ) 6.85-6.81 (1H,
m, H-6′), 6.69 (1H, m, H-5′), 6.69-6.64 (1H, m, H-2′), 5.48 (1H, s,
H-4), 5.46* (33% 2H, s, H-4*), 3.88* (30% 3H, s, OCH3*), 3.87 (3H,
s, OCH3), 3.54 (1H, 3-OH), 2.89-2.81 (2H, m, H-1), 2.57 (2H, m,
H-2), 2.49* (30% 2H, q, J ) 7.2 Hz, H-6*), 2.30 (2H, q, J ) 7.6 Hz,
H-6), 1.13 (3H, t, J ) 7.5 Hz, H-7), 1.04 (30% 3H, t, J ) 7.2 Hz,
H-7*) ppm. 13C NMR (100 MHz, CD3OD, internal standard TMS,
*diketo tautomer): δ ) 204.62* (C, C-5*), 203.55* (C, C-3*), 195.38
(C, C-5), 192.99 (C, C-3), 146.44* (C, C-3′*), 146.39 (C, C-3*),
143.99* (C, C-4′*), 143.97 (C, C-4′), 132,68 (C, C-1′), 132.48* (C,
C-1′*), 120.84 (CH, C-6′), 120.82* (CH, C-6′*), 114.34* (CH, C-5′*),
114.31 (CH, C-5′), 111.00* (CH, C-2′*), 110.95 (CH, C-2′), 98.75 (CH,
C-4), 57.12* (CH2, C-4*), 55.90* (CH3, OCH3*), 55.88 (CH3, O-CH3),
45.52* (CH2, C-2*), 40.45 (CH2, C-2), 37.13* (CH2, C-1*), 31.45 (CH2,
C-6), 31.37 (CH2, C-1), 29.16* (CH2, C-1*), 9.65 (CH3, C-7), 7.46*
(CH3, C-7*) ppm.
(1E)-1-(4-Hydroxy-3-methoxyphenyl)hept-1-ene-3,5-dione (11).
Vanillin (6.3 g, 42 mmol), 2,4-hexadione (14.1 g, 124 mmol) prepared
according to Ohta et al. (23), and boron trioxide (11.5 g, 165 mmol)
were mixed with N,N-dimethyl formamide (50 mL) and warmed to 90
°C. A solution of isobutylamine (0.041 g, 0.4 mmol) in N,N-dimethyl
formamide (38 mL) was added dropwise for 2 h. After stirring at 90

Short Chain Gingerdione Derivatives as Flavor Modifiers
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6659
(1E)-1-(4-Hydroxy-3-methoxyphenyl)hex-1-ene-3,5-dione (13). In
a flame dried apparatus, (1E)-1-(4-hydroxy-3-methoxyphenyl)hex-1-
en-3,5-dione (5) (2.36 g, 10 mmol) was suspended in dichloromethane
(20 mL) under nitrogen and cooled to -70 °C. To the solution, BBr3
in dichloromethane (3.76 g, 15 mmol in 10 mL) was added via a syringe
through a rubber septum, and the reaction mixture was maintained at
-65 to -70 °C for 4.5 h and then warmed up during the next 16 h. To
the solution, ice water (16 g) and ethyl acetate (50 mL) were added,
washed with brine, dried over Na2SO4, filtered, and evaporated to
dryness in vacuo. The raw hispolone (13) was purified with FC on
silica gel 60 using the eluent n-hexane/ethyl acetate 7:3 (v/v). Yield:
approximately 10 mg of yellow solid. HRMS: calcd. for C12H12O4
220.0736, found 220.0732. HPLC-MS (RP-18-Phase, APCI+): m/z )
220.97 (100%, [M + H]+), 219.28 (37%). 1H NMR (400 MHz, CDCl3,
internal standard TMS): δ ) 8.34 (1H, bs, OH), 7.48 (1H, d, J ) 15.9
Hz, H-1), 7.16 (1H, d, J ) 2 Hz, H-2′), 7.04 (1H, ddd, J ) 8.2 Hz, J
) 2.1 Hz, J ) 0.5 Hz, H-6′), 6.87 (1H, d, J ) 8.2 Hz, H-5′), 6.47 (1H,
d, J ) 15.8 Hz, H-2), 5.78 (1H, s, H-4), 2.11 (3H, s, H-6) ppm. 13C
NMR (100 MHz, CD3OD, internal standard TMS): δ ) 197.78 (C,
C-5), 179.36 (C, C-3), 148.57 (C, C-3′ or 4′), 146.38 (C, C-4′ or 3′),
140.91 (CH, C-1), 128.35 (C, C-1′), 122.42 (C-6′ or C-2), 120.72 (CH,
C-2 or C-6′), 116.41 (CH, C-2′ or C-5′), 115.07 (CH, C-5′ or C-2′),
101.14 (CH, C-4), 26.60 (CH3, C-6) ppm.
warmed to 90 °C under nitrogen for 3 h and subsequently refluxed for
2 h. The hot reaction mixture was poured on ice/water (600 g) and
stirred for 1 h. The O-benzylferulic acid (22) was filtered off, washed
with water, and dried at 40 °C/1 mbar for 12 h. Yield: 83.8 g (294
mmol, 97%); LCMS (RP-18 phase, ESI+): m/z ) 285 ([M + H]+,
100%). The protected ferulic acid 22 (26.4 g, 93 mmol) was dissolved
in toluene (300 mL), and at ambient temperature, thionylchloride (14.3
g, 120 mmol) was added. The mixture was stirred until dissolution for
6 h and subsequently warmed up to 50 °C for 1 h. The solvent and
excess thionylchloride were removed by evaporation in vacuo to yield
O-benzylferulic acid chloride (23). In a flame dried apparatus under
nitrogen, magnesium turnings (2.67 g) were placed into the flask, and
dry ethanol (15 mL), CCl4 (0.5 mL), and dry diethyl ether (100 mL)
were carefully added. The mixture was stirred at 20-23 °C for 4 h.
After cooling down to 0 °C, ethyl acetoacetate (13 g, 76 mmol) in dry
diethyl ether (30 mL) was added dropwise during 1 h. After cooling to
-5 °C, the acid chloride 23 dissolved in dry diethyl ether (100 mL)
and dry tetrahydrofurane (150 mL) was added dropwise over a period
of 1 h. The mixture was stirred for 12 h at 0-23 °C and was diluted
with ice cold 50% sulfuric acid (100 g). The organic phase was washed
with water, dried with Na2SO4, and filtered, and the filtrate was
evaporated to dryness to yield crude 24 (35 g, about 70% purity, 80%
yield based on ethyl acetoacetate). A portion was purified with flash
chromatography on silica gel 60 using the eluent n-hexane/ethyl acetate
3:1 (v/v). Yield: 7.6 g of a >90% fraction. LCMS (RP-18 phase, APCI-):
m/z ) 395, 29 ([M - H]-, 100%). Ethyl 2-acetyl-5-(4-benzyloxy-3-
methoxyphenyl)-3-oxopent-2-enoate (1.3 g, 3.3 mmol) (24) was dis-
solved in ethanol (20 mL), Pd-C (5%, wet) was added, and the mixture
was hydrogenated at ambient pressure and temperature with hydrogen.
After filtration, the filtrate was evaporated to dryness in vacuo, and
the crude product was purified with flash chromatography on silica gel
60 using the eluent n-hexane/ethyl acetate 2:1 (v/v). Yield: 0.42 g (purity
>95%, 1.4 mmol, 42%). LCMS (RP-18 phase, ESI+): m/z ) 309.6
([M + H]+, 100%). HRMS: calcd. for C16H20O6 308.1260, found
(1E)-1-(3,4-Diydroxyphenyl)hexa-3,5-dione (14). 1-(4-Hydroxy-
3-methoxyphenyl)hexan-3,5-dione (6) (2.39 g, 10 mmol) was treated
in the same way as described for hispolon (13). The raw dihydrohis-
polone (14) was purified with FC on silica gel 60 using the eluent
n-hexane/ethyl acetate 7:3 (v/v). Yield: 0.81 g (36%) of off-white solid
(GC DB1: >99%). MS (EI): m/z ) 222 (M+., 32%), 204 (2%), 164
(4%), 136 (29 ′%), 123 (100%′), 85 (39%), 43 (40%), 28 (79%). HRMS:
1
calcd. for C12H14O4 222.0892, found 222.0891. H NMR (400 MHz,
CDCl3, internal standard TMS, *diketo tautomer): δ ) 6.77 (1H, d, J
) 8.1 Hz, H-5′), 6.76* (20% 1H, d, J ) 8.0 Hz, H-5′), 6.69 (1H, d, J
) 2.0 Hz, H-2′), 6.67* (20% 1H, d, J ) 2.0 Hz, H-2′), 6.59 (1H, dm,
J ) 8.0 Hz, 2.0 Hz, H-6′), 6.57* (20% 1H, dm, J ) 8.1 Hz, J ) 2.0
Hz, H-6′), 5.49 (1H, s, H-4), 3.55 (30% 1H, s, H-4), 2.83-2.77 (m,
H-1 or H-2), 2.54 (2H, m, H-2 or H-1), 2.20* (20% 3H, m, J ) 0.5
Hz, H-6), 2.04 (3H, s, H-6) ppm. 13C NMR (100 MHz, CD3OD, internal
standard TMS, *diketo tautomer): δ ) 193.72 (C, C-3), 191.71 (C,
C-5), 143.77* (C, C-3′ or 4′), 143.73 (C, C-3′ or 4′), 142.19* (C, C-4′
or 3′), 142.10 (C, C-4′ or 3′), 133.42 (C, C-1′), 120.52 (CH, C-6′),
115.39* (C-2′ or C-5′), 115.34 (2 × CH, C-2′ and C-5′), 100.25 (CH,
C-4), 57.97* (CH2, C-4), 45.37* (CH2, C-2), 40.16 (CH2, C-2), 31.00*
(CH3, C-6), 30.94 (CH2, C-1), 28.79* (CH2, C-1), 24.89 (CH3, C-6)
ppm.
1
308.1272. H NMR (400 MHz, CDCl3, internal standard TMS): δ )
6.82 (1H, dd, J ) 7.9 Hz, J ) 0.3 Hz, H-5′), 6.71 (1H, d, J ) 1.9 Hz,
H-2′), 6.69 (1H, dd, J ) 7.9 Hz, J ) 2 Hz, H-6′), 5.59 (1H, s, 4′-OH),
4.25 (2H, q, J ) 7.2 Hz, Et-CH2), 3.86 (3H, d, J ) 0.2 Hz, O-CH3),
3.00-2.95 (2H, m, H-2), 2.91-2.86 (2H, m, H-1), 2.35 (3H, d, J )
0.2 Hz, H-6), 1.32 (3H, t, J ) 7.2 Hz, Et-CH3) ppm. 13C NMR (100
MHz, CD3OD, internal standard TMS): δ ) 198.09 (C, C-3), 195.61
(C, C-5), 167.15 (C, C-1′′), 146.45 (C, C-3′), 144.02 (C, C-4′), 132.62
(C, C-1′), 120.89 (CH, C-6′), 114.37 (CH, C-5′), 111.03 (CH, C-2′),
60.84 (CH2, Et-CH2), 55.87 (CH3, O-CH3), 39.93 (CH2, C-2), 31.31
(CH2, C-1), 25.62 (CH3, C-6), 14.22 (CH3, Et-CH3) ppm.
3-Oxobutyric acid N-(4-hydroxy-3-methoxyphenyl)methylamide
(17). Vanillylamine hydrochloride (0.99 g, 5.2 mmol) was suspended
in 1,4-dioxane (10 mL) and stirred with triethylamine (2 mL) for 2 h.
Ethyl acetoacetate (2.01 g, 15.5 mmol) was dissolved in 1,4-dioxane
(20 mL) and Chirazyme L2c2 (Boehringer Ingelheim, Germany) was
added. Under nitrogen atmosphere, the vanillylamine solution was added
portionwise (5 × 2 mL) to the ethyl acetoacetate solution at room
temperature during 5 h and stirred for a further 12 h. The precipitate
was removed by filtering and washing with dichloromethane, and
the combined filtrates were evaporated in vacuo to dryness. The oily
residue was recrystallized from ethyl acetate to yield 0.25 g (20%) of
pale white crystals. HPLC-MS (RP-18-Phase, APCI+): m/z ) 474.53
(22%, [2M]+), 237.85 (100%, [M]+). HRMS: calcd. for C12H15O4N
237.1001, found 237.1004. 1H NMR (400 MHz, CDCl3, internal
standard TMS): δ ) 7.20 (1H, bs, NH), 6.85 (1H, d, J ) 8.0 Hz, H-5′),
6.82 (1H, d, J ) 1.9 Hz, H-2′), 6.77 (1H, ddd, J ) 8.0 Hz, J ) 2.0 Hz,
J ) 0.6 Hz, H-6), 5.74 (1H, s, OH), 4.37 (2H, d, J ) 5.7 Hz, H-7′),
3.88 (3H, s, OCH3), 3.44 (2H, s, H-2), 2.26 (3H, m, J ) 0.4 Hz) ppm.
13C NMR (100 MHz, CD3OD, internal standard TMS): δ ) 204.53
(C, C-3), 165.34 (C, C-1), 146.74 (C, C-3′), 145.13 (C, C-4′), 129.83
(C, C-1′), 120.76 (CH, C-6′), 114.45 (CH, C-5′), 110.62 (CH, C-2′),
55.96 (CH3, O-CH3), 49.63 (CH2, C-2), 43.50 (CH2, C-7′), 31.09 (CH3,
C-4) ppm.
1,7-Bis(4-hydroxy-3-methoxyphenyl)-hepta-3,5-dione (20). Tet-
rahydrocurcumin (20) was obtained by hydrogenation of curcumin (1.5
g, 4.1 mmol) analogous to the described procedure for 6 and
subsequently purified with flash chromatography on silica gel using
the eluent n-hexane/ethyl acetate 1:1. Yield: 0.60 g (39%). LC-MS
(APCI+): m/z ) 389.8 (100%, [M + H2O]+). 1H NMR (400 MHz,
CDCl3, internal standard TMS): δ ) 6.83 (2H, m), 6.68 (4H, m), 5.48
(2H, OH), 5.42 (1H, s, H-4), 3.86 (3H, s, O-CH3), 2.85 (4H, m), 2.55
(4H, m) ppm. 13C NMR (100 MHz, CD3OD, internal standard TMS,
*diketo tautomer): δ ) 203.44* (C, C-3*, C-5*), 193.24 (C, C-3, C-5),
146.43 (C, C-3′), 143.96 (C, C-4′), 132.56 (C, C-1′), 132.34* (C, C-1′*),
120.81 (CH, C-6′), 114.35 (CH, C-5′), 111.02* (CH), 110.95 (CH, C-2′),
99.82 (CH, C-4), 57.65* (OCH3*), 55.87 (CH3, C-2), 45.52*, 40.38
(CH2), 31.31 (CH2, C-1) ppm.
Sensory Studies. All raw sensory data were analyzed using the
standard functions of Microsoft Excel 97. For the calculation of
statistical significance, Student’s matched pair test was used.
Dose Response Plots for Sucrose and Caffeine. Series of dilutions
of sucrose (0.29-83% by matter) and caffeine (100-28 340 mg kg-1
)
were presented to the panelists (trained, 25 and 20, respectively) in
the order of increasing concentration (unknown to the panelist) with
the advice to rate directly without backward tasting again. The rating
was given on an unstructured 15 cm scale from left (nothing) to right
(strongest taste effect). The test was performed twice, and the values
were averaged and recalculated to a scale from 1 to 10.
Ethyl 2-Acetyl-5-(4-hydroxy-3-methoxyphenyl)-3-oxopentanoate
(18). O-Benzylvanillin (72.9 g, 302 mmol) (21), malonic acid (62.7 g,
603 mmol), and piperidine (3 mL) in dry pyridine (165 mL) were

6660 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Ley et al.
Figure 3. Synthesis of gingerdione derivatives 5-14: (i) (a) B2O3, acetyl acetone, (b) Bu3B, EtOAc, (c) aldehyde, EtOAc, and (d) n-BuNH2/EtOAc; (ii)
Pd-C, H2, EtOH; (iii) BBr3, CH2Cl2, -70 °C, room temperature. * alternative method: (i) (a) B2O3, 3,5-heptadienone, aldehyde, isobutylamine, DMF,
90 °C.
Bitter Masking and Sweet Enhancing ActiVity. For screening and
dose response studies of taste modulating effects, the test compounds
were added directly to an aqueous solution of the appropriate bitter or
sweet compound. Occasionally, the mixture was treated for several
minutes in an ultrasonic bath to improve the dissolution process.
Panelists (healthy adults, no tasting problems known) were trained on
caffeine as a bitter standard and sucrose as sweet standard. Studies
were performed in the morning hours 1 to 2 h after breakfast, during
which time they were not allowed to drink black or green tea or coffee
due to adaptation to caffeine; only one test per day was performed. A
minimum of 10 testers were used in the descriptive test. For all
experiments, the test solutions were coded, and in the case of color or
cloudiness, the cups were covered using an aluminum foil. Panelists
were advised to test randomly mixed samples in the given order by
the sip and spit method. For bitter masking activity, the bitter compound
Figure 4. Synthesis of gingerdione derivative 18: (i) malonic acid, pyridine,
was tested only in one concentration (given in the tables, e.g., 500 mg
piperidine (97%); (ii) SOCl2, toluene (quant.); (iii) Mg/EtOH/THF/DEE
kg-1 for caffeine) showing sufficient bitterness that was determined in
(80%); (iv) H2, Pd-C, EtOH (42%).
preliminary tests. For screening of sweet enhancing activity, 5% sucrose
solution was used. For screening purposes, the taste quality was
en-3,5-dione was found only in amounts <5% and was separated
by flash chromatography.
absolutely rated on a scale of 1 (no taste) to 10 (very strong taste).
Relative flavor modification activity (in %) was calculated by averaging
the intensity ratings for each the test and blank solution and then using
the following formula: averaged intensity ratingtest/averaged intensity
ratingblank - 1.
Determination of Intrinsic Sweetness. For determination of intrinsic
sweetness, an aqueous solution of the test compound in the appropriate
concentration was compared to a series of dilutions of sucrose (0, 0.25,
0.5, 0.75, 1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, and 10%). The values of all
panelists for each concentration of the test compound were averaged.
The plots were given as percent sucrose equivalents.
The gingerdiones 6, 8, 10, and 12 were easily obtained by
Pd catalyzed hydrogenation at ambient pressure. The reaction
mixtures contained as main components the final products and
as side products the appropriate gingerols in amounts <5%.
Hispolon (13) and dihydrohispolon (14) were synthesized using
the well-established cleavage of the methyl ether moiety of
[2]-dehydrogingerdione (5) and [2]-gingerdione (6), respectively,
with BBr3 in dichloromethane (Figure 3). The reaction yielded
only very small amounts of 5 due to interference of the
conditions with the unsaturated skeleton and due to losses during
the workup. Tetrahydrocurcumin (20) was prepared starting from
commercial curcumin (19) using the same hydrogenation
conditions; the analytical data corresponded to known data from
the literature (28).
Acetoacetic acid vanillylamide (17) was prepared via ami-
dation of ethyl acetoacetate with vanillylamine in 1,4-dioxane/
triethylamine using Candida antarctica lipase (Chirazym c2 L2)
as catalyst and subsequent recrystallization. Amidation reactions
using lipases are a known conversion (29), but to our knowledge,
the reaction with ethyl acetoacetate and benzyl amines was never
described.
RESULTS AND DISCUSSION
Synthesis. The dehydrogingerdiones 5, 7, 9, and 11 were
synthesized using the known procedure starting from an excess
of an alka-3,5-dione-boron complex, which is reacted with the
appropriate benzaldehyde with the aid of triethylborate under
basic catalysis (24). When the aldehyde was used in more than
an equimolar ratio, the main product is the appropriate curcumin
derivative. The remaining amounts of the bis-coupling products
were easily separated by crystallization and if necessary by flash
chromatography. Syntheses of [2]- and [3]-dehydrogingerdione
(5, 11) as well as for [2]-isogingerdione (7) were described
earlier (25) as well as for the p-hydroxy derivative 9 (26). In
the case of 11, we have chosen an improved variation of the
method using DMF as the solvent without the use of triethyl
borate as described for an improved synthesis of curcumin (27).
The reaction was much more straightforward and much easier
to clean up. Fortunately, the reaction yielded as main component
the intended regioisomer, and the alternative condensation
product (1E)-2-methyl-1-(4-hydroxy-3-methoxyphenyl)hex-1-
The branched gingerdione ethyl-5-(4-hydroxy-3-methoxyphe-
nyl)-2-(1-oxoethyl)-3-oxopentanoate (18) was synthesized as
shown in Figure 4 analogously to ethyl-5-(4-hydroxyphenyl)-
2-(1-oxoethyl)-3-oxopentanoate, as described by Doherty (30).
O-Benzyl vanillin (21) was converted to O-benzyl ferulic (22)
acid via malonic ester condensation and reacted with thionyl-
chloride to yield the appropriate acid chloride 23. The latter
was coupled to ethyl acetoacetate using magnesium in diethyl

Short Chain Gingerdione Derivatives as Flavor Modifiers
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6661
Figure 5. Sweetness of 0.29-83% aqueous sucrose solutions (25
panelists, free choice on a 15 cm unstructured scale, recalculation to
scale 1-10).
Figure 6. Bitterness of 100-28 340 mg kg-1 aqueous caffeine solution
(20 panelists, free choice on a 15 cm unstructured scale, recalculation to
scale 1-10).
ether/THF/ethanol to give 24 and subsequently hydrogenated
to yield the final product 18.
Taste modulating effects lower than 15% were considered
only as weak from a practical point of view and are in nearly
all cases statistically not significant using the low number of
available panelists (10-20). Effects lower than 10% were not
considered as valid. In most cases, masking or enhancing effects
larger than 25 or 30% are statistically significant and such
molecules were evaluated as potential taste modulators.
All compounds were purified to at least 95% and character-
ized using (if applicable) GC-MS, LC-MS, and HRMS in the
case of new or poorly described products and NMR techniques
prior to taste tests. In all cases, the beta-keto system showed
tautomerization, and in the solvent used for recording NMR
spectra, the main tautomers (>80%) were the enol-isomers.
Screening for Taste Modulating Effects. In our efforts to
screen for taste modulating compounds, we used double blinded
duo testing. The methods are generally state of the art and used
for the screening of taste influencing substances since their
introduction by Jellinek in 1966 (31). Important criteria for
reliable results is a trained panel with sufficient participants,
randomized and blinded sampling, and preferably only one
tasting session per day, which is best performed in the morning.
The panelists record their bitterness impression on a hedonic
scale. Some other working groups have developed the half-site
tongue test, which may be preferred for very small sample
volumes (32). We did not use the latter method because in
preliminary tasting sessions the masking and enhancing effects
of known taste modulators could not be fully reproduced due
to the very low volumes. An alternative test, the half-mouth
method (33) using higher amounts of test solutions was not used
due to the sophisticated panelist training necessary for reliable
results.
Prior to the screening for sweetness enhancing effects, we
tested the sweetness/concentration relationship of neat sucrose
solution. The panelists were asked to rate the sweetness of
different concentrations of sucrose on an unstructured scale of
15 cm, which was recalculated to a rating from 1 (weak) to 10
(very strong). Despite the fact that most sweet applications
(mainly beverages, ice cream etc.) contain between 8 and 15%
sucrose, we chose for screening a lower concentration of 5%
sucrose solution as test medium because then changes
in sweetness could most easily be detected (Figure 5). A 5%
sucrose solution is about 25% less sweet compared to a 6%
sucrose solution when compared directly in the duo test.
Therefore, for real applications, a sweetness improving flavor
compound should exhibit an enhancing effect of at least 25%.
Similar to sweetness rating, a dose rating plot was determined
using a series of dilutions of caffeine in water (cf. Figure 6).
In contrast to sugar as test medium, the acceptance of a
1000-2000 mg kg-1caffeine solution is very low. Therefore,
we decided to perform the tests using 500 mg kg-1 caffeine.
The ratings during a single masking trial at 500 mg kg-1 caffeine
showed generally higher values when they were compared to
the ratings of the same caffeine concentration in the series of
dilutions. The differences may be caused by the strong lingering
and adaptation effects of caffeine during the dose-response plot
tastings.
The most simplified structural elements of homoeriodictyol,
dehydrozingerone (3), and zingerone (4) showed only moderate
taste modulation effects (cf. Table 1). In the latter case, the
concentration used was much lower than for most other test
compounds due to the strong intrinsic flavor of 4. [2]-Dehy-
drogingerdione (5) and to a much higher extend [2]-gingerdione
(6) showed a pronounced, and in the case of the latter, highly
significant masking effect against caffeine bitterness. The sweet
enhancing effect on a 5% sucrose solution of both molecules
was somewhat weak. Both molecules showed no detectable
intrinsic sweetness, and therefore, we conclude that they act as
real bitter masking compounds. Exchange of the vanillyl against
an isovanillyl aromatic moiety leads to a change in the activity
pattern; whereas the masking activity was lower compared to
5, the sweetness enhancing activity of the [2]-isogingerdione
was much higher and highly significant in contrast to the
regioisomer. The substitution pattern on the aromatic ring seems
to be crucial for activity: the simple p-hydroxyphenyl derivatives
9 and 10 showed much lower taste modulation effects as
compared to their higher substituted counterparts. Elongation
of the alky chain leads to lower efficacy for masking activity:
the naturally occurring [3]-dehydrogingerdione (11) showed only
weak masking and no sweetness enhancing effect. The corre-
sponding [3]-gingerdione (12) was remarkably effective against
caffeine but weaker than the lower homologous 6. The sweetness
enhancing effect of 12 was higher than for 6 but much lower
compared to the [2]-isogingerdione (8). Because of the low
amounts available, hispolon (13) was only tested by qualitative
judgment in sucrose and caffeine solutions; at 50 mg kg-1, it
exhibited only a weak bitter enhancing effect and was not able
to increase sweetness. Dihydrohispolon (14) demonstrated at
100 mg kg-1 a pronounced dull taste/flavor and was therefore
not tested for bitter masking activity. It showed only a weak
sweet enhancing effect. The long chain relatives [8]- and [10]-
gingerol (15, 16) showed no activity at all. The test concentration
of the latter two compounds was 10 times lower as compared
to most of the other molecules due to the intrinsic flavor, which
was yet remarkably pungent at 10 mg kg-1. N-Vanillyl
acetoacetamide 17, the aza-analogue of 6, as well as the
branched variation of 6, the 2-ethylcarboxy [2]-gingerdione (18),
showed only moderate and nonsignificant taste modulation

6662 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Ley et al.
Table 1. Evaluation of Taste Modulation Effects of Compounds Compared to Homoeriodictyol 1 against 500 mg kg-1 Caffeine and 5% Sucrose,
Respectivelya
bitter masking (500 mg kg-1 caffeine)
compd panelists (all/masking)c reduction of bitter rating panelists (all/enhancing)d enhancing of sweet rating
sweet enhancing (5% sucrose)
profile (100 mg kg-1 in 5% sucrose)
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
10/10
16/8
15/8
43%* (p < 0.05)
10%
15/9
16/1
18%
-15%
n.d.
15%
11%
-4%
27%* (p < 0.005)
2%
10%
weak, sweet, vanillic, phenolic
sweet, balsamic, numbing, neutral
0.7% (1 mg kg-1
21% (p < 0.07)
35%* (p < 0.005)
23%* (p < 0.05)
17%
)
spicy, clove, smoky (1 mg kg-1
neutral, drying
weak, cream, sweet, vanillic
)
12/9
16/10
15/9
16/7
16/13
16/5
16/9
17/15
16/12
15/11
16/9
16/8
16/9
weak, dry dusty, balsamic, tingling
weak, phenolic, smoky, medicinal
weak, soapy, sweet, dry-dusty, balsamic
weak, raspberry, sweet, mouth feel
weak, sweet, vanillic, ginger
weak, vanillic, smoke, phenolic
mouth feel, astringent
6%
17%
18%
27%* (p < 0.05)
more bitter (50 mg kg-1 b
14/6
16/11
0%
19% (p < 0.07)
16/13
)
no difference (50 mg kg-1 b
8%
)
n.d.
16/7
16/9
dull, anthranilate
roasty, coffee (10 mg kg-1
)
16/9
16/6
15/10
16/6
16/7
11/6
-4% (10 mg kg-1
-3% (10 mg kg-1
12%
)
)
16% (5 mg kg-1
n.d.
)
burning, numbing (10 mg kg-1
neutral
woody, card board
)
16/8
16/10
16/3
15/6
10%
11%
-5%
-6% (1 mg kg-1
9%
)
-2% (1 mg kg-1
-1%
)
sweet, bitter, caramelic (1 mg kg-1
)
fatty, sweet almond, fruity, mouth feel, long lasting
a Test concentration 100 mg kg-1; * significant (p < 0.05); n.d. ) not determined. b Only qualitative judgment due to low amounts available. c Ratio of number of all
panelists against number of panelists who rated the bitterness of test solution lower than standard solution. d Ratio of number of all panelists against number of panelists
who rated the sweetness of test solution higher than standard solution.
activities. The related diarylheptanoids curcumin (19, only 1
mg kg-1 tested due to the strong flavor profile) and tetrahy-
drocurcumin (20) showed no taste modulation effects at all.
Evaluation of [2]-Gingerdione (6) and [3]-Gingerdione (12)
as Bitter Masking Compounds. Prior to further evaluation for
gingerdione 6, an Ames test (Salmonella typhimorum reversed
mutation assay) with/without metabolic activation was per-
formed according to OECD guidelines (34). Gingerdione 6
showed no mutagenic effect in the Salmonella typhimorum
strains TA98, TA100, TA102, TA1535, and TA1537 each
carried out with/without metabolic activation. Because of the
structural similarity, the isogingerdione 8 and gingerdione 12
were not tested using the Ames protocol.
In Figure 7, gingerdiones 6 and 12 exhibit clear dose-
response behavior against the bitterness of caffeine. Analogously
to homoeriodictyol (17), the masking effect reaches a maximum
activity at a 100 mg kg-1 level. As shown in Table 2, the
activity of the gingerdiones against the bitterness of caffeine
and quinine is quite similar, whereas both compounds in contrast
to homoeriodictyol, do not reduce salicin bitterness at the tested
concentration of 100 mg kg-1. Gingerdione 6 was not active
against bitter peptides, whereas gingerdione 12 exhibited a weak,
nonsignificant masking effect against bitterness of KCl.
In general, the masking effect of the gingerdiones against
caffeine and quinine are quite comparable to homoeriodictyol
and 2,4-dihydrobenzoic acid N-vanillylamide (2) (15), whereas
the tested activity patterns of 6 and 12 show some remarkable
differences in respect to salicin. This is of interest for elucidation
of the hitherto unknown masking mechanism of homoeriodictyol
derivatives. In our opinion, these structural classes may act as
antagonists for some but not the whole set of bitter taste
receptors.
Evaluation of [2]-Isogingerdione (8) as a Sweetness
Enhancer. The isovanillyl moiety is known as a typical
structural element occurring in high potency sweeteners (35, 36).
Therefore, we evaluated the intrinsic sweetness of a series of
dilutions of isogingerdione 8 to exclude a simple additive
sweetening effect. For comparison, the sweetness of sucralose
was calculated according to DuBois et al. (37) (Figure 8) and
Figure 7. (A) [2]-Gingerdione (6) and (B) [3]-gingerdione (12) dose/activity
plot for bitterness masking of a 500 mg kg-1 caffeine solution determined
by paired duo test (trained panel, n ) 16). * significant, p < 0.05.
plotted along with the measured sweetness of solutions contain-
ing 8. The intrinsic sweetness of 8, even at 250 mg kg-1, was
at about 1% sucrose equivalent, and it was 0.58% sucrose
equivalent at 100 mg kg-1. Most panelists did not perceive a
0.58% sucrose solution as sweet (averaged rating on the 1-10
scale at 1.5).
To determine synergistic activity of 8, we have compared
calculated and experimental sweetness ratings of 5% sucrose
solutions containing various amounts of isogingerdione 8
(Figure 9). For calculation reasons, we have normalized the
absolute sweet ratings (1-10) to sucrose equivalents (SE) using

Short Chain Gingerdione Derivatives as Flavor Modifiers
J. Agric. Food Chem., Vol. 56, No. 15, 2008 6663
Table 2. Bitter Masking Effect of 100 pp mg kg-1 [2]-Gingerdione (6) and
[3]-Gingerdione (12) against Various Bitter Compoundsa
[2]-gingerdione (6)
[3]-gingerdione (12)
panelists reduction of panelists reduction of
(all/masking)b bitter rating (all/masking)b bitter rating
bitter compd
caffeine, 500 mg kg-1
quinine, 5 mg kg-1
salicine, 250 mg kg-1
H-Leu-Trp-OH, 2000
mg kg-1
17/15
16/10
16/6
35%*
21%*
3%
16/13
15/10
16/7
27%*
22%
2%
16/6
1%
n.d.
KCl, 0.5%
n.d.
15/8
5%
a Parameters cf. Table 1. * significant, p < 0.05. n.d. ) not determined. b Ratio
of number of all panelists against number of panelists who rated the bitterness of
test solution lower than standard solution.
Figure 10. Change in sweetness ratings of a 10% sucrose solution against
an 8% sucrose solution (blank), an 8% sucrose solution containing 0.01%
stevioside, and an 8% sucrose solution containing gingerdione 8
determined via a paired duo test (trained panel, n ) 15); significance
was determined using Student’s t test.
More important for application is the comparison of full sugar
applications with sugar reduced variants treated with a sweetness
enhancer. Therefore, we performed a paired duo test using a
10% sucrose solution against an 8% sucrose solution with/
without test compound 8 (Figure 10). For comparison, stevio-
side as a typical high intensity sweetener was used. Surprisingly,
only 8 was able to increase the perceived sweetness of the 8%
sucrose solution to a level of 10%.
By starting from the structure of the bitter masking compound
homoeriodictyol, it is possible to reduce the complexity of the
flavanone skeleton without reduction of the masking effect
against caffeine and quinine, as shown for the short chain
gingerdiones 6 and 12. In contrast, the ability to reduce salicin
bitterness is not evident. Further modifications of the structural
motifs (chain elongation, different substitution pattern, reducing
flexibility) cause activity loss. Interestingly, the isogingerdione
8 is able to induce sweetness enhancing effects in model
solutions without exhibiting an intrinsic sweet taste.
Figure 8. Intrinsic sweetness of [2]-Isogingerdione (8) determined by direct
comparison of a series of dilutions of test compounds against standard
sucrose solutions (0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, and
10%) (trained panel, n ) 23). Sucralose sweetness was calculated
according to DuBois et al. (37), R ) 13c(sucralose)1.4/(1101.4
+
c(sucralose)1.4). The units of R are % sucrose equivalents; the units of c
are mg kg-1
.
ACKNOWLEDGMENT
We would like to thank Katharina Reichelt and Franziska
Benedix for compilation of the dose-response plots for sucrose
and caffeine and Martina Herrmann and Holger Joppe for
isolation of the gingerols. Special thanks to Debbie Kennison
for proof reading of the manuscript.
LITERATURE CITED
Figure 9. Dose-response plot of sweetness enhancing activity of
isogingerdione 8 in 5% sucrose solution compared to the calculated sum
of sucrose equivalents of the ratings for a 5% sucrose solution (calculated
using the plot in Figure 5) and the intrinsic sweetness of 8 (trained panel,
n ) 15).
(1) Drewnowski, A.; Gomez-Carneros, C. Bitter taste, phytonutrients,
and the consumer: a review. Am. J. Clin. Nutr. 2000, 72, 1424–
1435.
(2) Sohi, H.; Sultana, Y.; Khar, R. K. Taste masking technologies in
oral pharmaceuticals: recent developments and approaches. Drug
DeV. Ind. Pharm. 2004, 30, 429–448.
(3) Katsugari, Y.; Sugiura, Y.; Lee, C.; Otsuji, K.; Kurihara, K.
Selective inhibition of bitter taste of various drugs by lipoprotein.
Pharm. Res. 1995, 12, 658–662.
the plot in Figure 5. For each concentration, the SE of the 5%
sucrose solutions (Figure 9, solid triangles), the sum of the SE
of the 5% sucrose solution, the appropriate intrinsic SE of 8
(Figure 9, open cirles), and the measured SE of the test solution
containing 5% sucrose and 8 (Figure 9, solid circle) were
plotted. At 10, 20, 50, and 100 mg kg-1, respectively, the
calculated synergistic effect was 5, 14, 13, and 23%. The effect
at 100 mg kg-1 is evident. As a result, we conclude that an
amplification of sweetness response to sucrose can be induced
not only by low amounts of high intensity sweeteners but also
by very weak sweet molecules.
(4) Sonoda, C.; Sakai, K.; Murase, K. Bitternis relieving agent. WO
00 21,390, 2000.
(5) Hayashi, N.; Ujihara, T.; Kohata, K. Reduction of catechin
astringency by the complexation of gallate-type catechins with
pectin. Biosci., Biotechnol., Biochem. 2005, 69, 1306–1310.
(6) Binello, A.; Cravotto, G.; Nano, G. M.; Spagliardi, P. Synthesis
of chitosan-cyclodextrin adducts and evaluation of their bitter-
masking properties. FlaVour Fragrance J. 2004, 19, 394–400.
(7) Busto, M. D.; Mezaa, V.; Ortega, N.; Perez-Mateos, M. Im-
mobilization of naringinase from Aspergillus niger CECT 2088

6664 J. Agric. Food Chem., Vol. 56, No. 15, 2008
Ley et al.
in poly(vinyl alcohol) cryogels for the debittering of juices. Food
Chem. 2007, 104, 1177–1182.
(8) Psczola, D. Ingredientssflavor enhancement: taking the mask off.
(23) Ohta, H.; Ozaki, K.; Tsuchihashi, G.-I. Regio- and enantio-
selective reduction of R,γ-diketones by fermenting baker’s yeast.
Agric. Biol. Chem. 1986, 50, 2499–2502.
Food Technol. 2004, 58, 56–69.
(9) Ley, J. P. Bitter masking bitter Ttaste by molecules. Chem.
Percept. 2008, 1, 58–77.
(24) Masuda, T.; Matsumura, H.; Oyama, A.; Takeda, Y.; Jitoe, A.;
Kida, A.; Hidaka, K. Synthesis of (()-cassumins A and B, new
curcuminoid antioxidants having protective activity of the living
cell against oxidative damage. J. Nat. Prod. 1998, 61, 609–613.
(25) Belliotti, T. R.; Connor, D. T.; Flynn, D. L.; Kostlan, C. R.; Nies,
D. E. Preparation of novel styrylpyrazoles, styrylisoxazols, and
analogs as 5-lipoxygenase inhibitors. EP 245,825, 1987.
(26) Mann, G.; Beyer, L.; Arrieta, A. Preparation of unsymmetrical
curcuminoids. Z. Chem. 1987, 27, 172–173.
(10) Salemme, F. R.; Barndt, R. Compositions and methods for
producing a salty taste in foods or beverages. U.S. Patent 2006
286,276, 2006.
(11) Soldo, T.; Hofmann, T. Application of hydrophilic interaction
liquid chromatography/comparative taste dilution analysis for
identification of a bitter inhibitor by a combinatorial approach
based on Maillard reaction chemistry. J. Agric. Food Chem. 2005,
53, 9165–9171.
(27) Wegner, C. Method for production of curcumin. WO 2006
089,894, 2006.
(12) Duffy, V. B.; Sigman-Grant, M.; Powers, M. A.; Elmore, D.;
Myers, E. F.; Quagliani, D.; Spano, M.; Stitzel, K. F.; Taylor, S.;
Earl, R.; Connor, S. Position of the American dietetic association:
use of nutritive and nonnutritive sweeteners. J. Am. Diet. Assoc.
2004, 104, 255–275.
(13) Ottinger, H.; Soldo, T.; Hoffmann, T. Discovery and structure
determination of a novel Maillard-derived sweetness enhancer by
application of the comparative taste dilution analysis (cTDA). J.
Agric. Food Chem. 2003, 51, 1035–1041.
(14) Labbe, D.; Rytz, A.; Morgenegg, C.; Ali, S.; Martin, N.
Subthreshold olfactory stimulation can enhance Ssweetness. Chem.
Sens. 2007, 32, 205–214.
(15) Ley, J. P.; Blings, M.; Paetz, S.; Krammer, G.; Bertram, H.-J.
New bitter masking compounds: hydroxylated benzoic acid amides
of aromatic amines as structural analogues of homoeriodictyol.
J. Agric. Food Chem. 2006, 54, 8574–8579.
(16) Ley, J. P.; Kindel, G.; Paetz, S.; Krammer, G. Hydroxydeoxy-
benzoins and the use thereof to mask a bitter taste. WO 2006
106,023, 2006.
(17) Ley, J. P.; Krammer, G.; Reinders, G.; Gatfield, I.-L.; Bertram,
H.-J. Evaluation of bitter masking flavanones from Herba Santa
(Eriodictyon californicum (H. & A.) Torr., Hydrophyllaceae). J.
Agric. Food Chem. 2005, 53, 6061–6066.
(18) Ley, J. P.; Paetz, S.; Kindel, G.; Freiherr, K.; Krammer, G. E.
Enhancers for sweet taste from the world of non-volatiles. In
Sweetness and SweetenerssBiology, Chemistry and Psychophys-
ics; Weerasinghe, D. K., DuBois, G., Eds.; Oxford University
Press: New York, 2008; pp 400-409.
(28) Osawa, T.; Sugiyama, Y.; Inayoshi, M.; Kawakishi, S. Antioxi-
dative activity of tetrahydrocurcuminoids. Biosci., Biotech., Bio-
chem. 1995, 59, 1609–1612.
(29) de Zoete, M. C.; Kock-van Dalen, A. C.; van Rantwijk, F.;
Sheldon, R. A. Lipase-catalyzed ammoniolysis of lipids. A facile
synthesis of fatty acid amides. J. Mol. Catal. B: Enzym. 1996, 2,
141–145.
(30) Doherty, D. G. The hydrolysis of carbon-carbon bonds by
R-chymotrypsin. J. Am. Chem. Soc. 1955, 77, 4887–4892.
(31) Jellinek, G. Moderne Verfahren der sensorischen analyse. Planta
Med. 1966, 14, 21–33.
(32) Shikata, H.; McMahon, D. B. T.; Breslin, P. A. Psychophysics
of taste lateralization on anterior tongue. Percept. Psychophys.
2000, 62, 684–694.
(33) Galopin C.; Tigani, L.; Furrer, S. Structure property relationship
of chemosensates. In Structure Property Relationship of Chemosen-
sates; Hofmann, T., Rothe, M., Schieberle, P., Eds.; Deutsche
Forschungsanstalt fu¨r Lebensmittelchemie: Garching, Germany,
2005; pp 61-67.
(34) Final Report; No. 18432/18/04; Laboratory of Pharmacology and
Toxicology GmbH & Co. KG: Hamburg, Germany, 2007.
(35) Arnoldi, A.; Bassoli, A.; Merlini, L. Progress in isovanillyl sweet
compounds. Food Chem. 1996, 56, 247–253.
(36) Bassoli, A.; Drew, M. G. B.; Hattotuwagama, C. K.; Merlini, L.;
Morini, G.; Widen, G. R. W. Quantitative structure-activty
relationships of sweet isovanillyl derivatives. Quant. Struct.-Act.
Relat. 2001, 20, 3–16.
(37) DuBois, G. E.; Walters, D. E.; Schiffman, S. S.; Warwick, Z. S.;
Booth, B. J.; Pecore, S. D.; Gibes, K.; Carr, B. T.; Brands, L. M.
A systematic study of cxoncentration-response relationships of
sweeteners. In Sweeteners: DiscoVery, Molecular Design, and
Chemoreception; Walters, D. E., Orthoefer, F. T., DuBois, G. E.,
Eds.; American Chemical Society: Washington, DC, 1991; pp
261-276.
(19) The number in the square brackets denotes the chain length of
the residue added to the parent compound zingerone (4). Kikuzaki,
H. Ginger for drug and spice purposes. Herbs, Bot. Teas 2000,
75–105.
(20) Jolad, S. D.; Lantz, R. C.; Solyo, A. M.; Chen, G. J.; Bates, R. B.;
Timmermann, B. N. Fresh organically grown ginger (Zingiber
officinale): composition and effects on LPS-induced PGE2 produc-
tion. Phytochemistry 2004, 65, 1937–1954.
(21) Jolad, S. D.; Lantz, R. C.; Chen, G. J.; Bates, R. B.; Timmermann,
B. N. Commercially processed dry ginger (Zingiber officinale):
composition and effects on LPS-stimulated PGE2 production.
Phytochemistry 2005, 66, 1614–1635.
(22) Ali, N. A. A.; Jansen, R.; Pilgrim, H.; Liberra, K.; Lindequist, U.
Hispolon, a yellow pigment from Inonotus hispidus. Phytochem-
istry 1996, 41, 927–929.
Received for review March 3, 2008. Revised manuscript received May
5, 2008. Accepted May 10, 2008. Parts of the study were presented on
the ACS National meeting, Atlanta, March 2006, and the eighth
Wartburg Symposium on Flavour Chemistry & Biology, February 28
to March 2, 2007, Eisenach, Germany.
JF8006536