In Vivo Studies on the Metabolism of the Monoterpene Pulegone in Humans Using the Metabolism of Ingestion-Correlated Amounts (MICA) Approach: Explanation for the Toxicity Differences between (S)-(-)- and (R)-(+)-Pulegone

Article abstract of DOI:10.1021/jf034702u

The major in vivo metabolites of (S)-(-)-pulegone in humans using a metabolism of ingestion-correlated amounts (MICA) experiment were newly identified as 2-(2-hydroxy-1-methylethyl)-5-methylcyclohexanone (8-hydroxymenthone, M1), 3-hydroxy-3-methyl-6-(1-methylethyl)cyclohexanone (1-hydroxymenthone, M2), 3-methyl-6-(1-methylethyl)cyclohexanol (menthol), and E-2-(2-hydroxy-1-methylethylidene)-5-methylcyclohexanone (10-hydroxypulegone, M4) on the basis of mass spectrometric analysis in combination with syntheses and NMR experiments. Minor metabolites were be identified as 3-methyl-6-(1-methylethyl)-2-cyclohexenone (piperitone, M5) and α,α,4-trimethyl-1-cyclohexene-1-methanol (3-p-menthen-8-ol, M6). Menthofuran was not a major metabolite of pulegone and is most probably an artifact formed during workup from known (M4) and/or unknown precursors. The differences in toxicity between (S)-(-)- and (R)-(+)-pulegone can be explained by the strongly diminished ability for enzymatic reduction of the double bond in (R)-(+)-pulegone. This might lead to further oxidative metabolism of 10-hydroxypulegone (M4) and the formation of further currently undetected metabolites that might account for the observed hepatotoxic and pneumotoxic activity in humans.

Full text of DOI:10.1021/jf034702u

J. Agric. Food Chem. 2003, 51, 6589−6597 6589
In Vivo Studies on the Metabolism of the Monoterpene
Pulegone in Humans Using the Metabolism of
Ingestion-Correlated Amounts (MICA) Approach: Explanation
for the Toxicity Differences between (S)-(−)- and
(R)-(+)-Pulegone
WOLFGANG ENGEL*
Deutsche Forschungsanstalt fu¨r Lebensmittelchemie, Lichtenbergstrasse 4,
D-85748 Garching, Germany
The major in vivo metabolites of (S)-(-)-pulegone in humans using a metabolism of ingestion-
correlated amounts (MICA) experiment were newly identified as 2-(2-hydroxy-1-methylethyl)-5-
methylcyclohexanone (8-hydroxymenthone, M1), 3-hydroxy-3-methyl-6-(1-methylethyl)cyclohexanone
(1-hydroxymenthone, M2), 3-methyl-6-(1-methylethyl)cyclohexanol (menthol), and E-2-(2-hydroxy-
1-methylethylidene)-5-methylcyclohexanone (10-hydroxypulegone, M4) on the basis of mass spec-
trometric analysis in combination with syntheses and NMR experiments. Minor metabolites were be
identified as 3-methyl-6-(1-methylethyl)-2-cyclohexenone (piperitone, M5) and R,R,4-trimethyl-1-
cyclohexene-1-methanol (3-p-menthen-8-ol, M6). Menthofuran was not a major metabolite of pulegone
and is most probably an artifact formed during workup from known (M4) and/or unknown precursors.
The differences in toxicity between (S)-(-)- and (R)-(+)-pulegone can be explained by the strongly
diminished ability for enzymatic reduction of the double bond in (R)-(+)-pulegone. This might lead to
further oxidative metabolism of 10-hydroxypulegone (M4) and the formation of further currently
undetected metabolites that might account for the observed hepatotoxic and pneumotoxic activity in
humans.
KEYWORDS: Metabolism of ingestion correlated amounts (MICA); urinary metabolites; pulegone; menthol;
piperitone; 2-(2-hydroxy-1-methylethyl)-5-methylcyclohexanone; 3-hydroxy-3-methyl-6-(1-methylethyl)-
cyclohexanone; 3-methyl-6-(1-methylethyl)cyclohexanol; E-2-(2-hydroxy-1-methylethylidene)-5-methylcy-
clohexanone; 3-methyl-6-(1-methylethyl)-2-cyclohexenone; r,r,4-trimethyl-1-cyclohexene-1-methanol
INTRODUCTION
(R)-(+)-Pulegone (Figure 1) is a constituent of peppermint
and pennyroyal oil. Additionally, it occurs in many other foods
such as oregano, beans, and tea. It is widely used as a flavor
constituent having Generally Recognized As Safe (GRA) status
in the United States. In the European Union the following
maximum levels for pulegone in foodstuffs have been set: 25
mg/kg in foodstuffs in general; 100 mg/kg in beverages; 250
mg/kg in peppermint- or mint-flavored beverages; and up to
350 mg/kg in mint confectionery. Despite its use in flavorings,
there are severe toxic effects of pulegone at large doses (1, 2),
very recently discussed by the Scientific Committee of the
European Commission (3). The other enantiomer, (S)-(-)-
pulegone (Figure 1) is found rarely in essential oils.
Figure 1. Structures of (S)-(−)-pulegone and (R)-(+)-pulegone.
Menthofuran was very early identified as a metabolite of
pulegone formed after ingestion of large amounts of pennyroyal
oil (4), and it is generally accepted that the hepatotoxicity of
pulegone is due, at least in part, to this metabolite. There have
been several studies of (R)-(+)- and (S)-(-)-pulegone metabo-
lism using rats as test animals (5-7, 9, 10) and in vitro
experiments with rat liver microsomes (8, 9, 11) and human
cytochrome P450 enzymes (12). Additionally, very recently the
focus was extended to the role of the enzyme glutathione
There have been many investigations regarding the toxicity
of pulegone, especially of the (R)-(+)-enantiomer, which seems
to be much more toxic compared to the (S)-(-)-enantiomer.
* Telephone 0049 (0)89 28913261; fax 0049 (0)89 28914183; e-mail
wolfgang.engel@lrz.tum.de.
10.1021/jf034702u CCC: $25.00 © 2003 American Chemical Society
Published on Web 09/19/2003

6590 J. Agric. Food Chem., Vol. 51, No. 22, 2003
Engel
mmol) in CH2Cl2 (10 mL) was added a solution of 9-acetoxy-p-
menthan-3-ol (∼1 mmol) in CH2Cl2 (50 mL). The mixture was stirred
overnight at room temperature. The solution was concentrated to 5 mL
using a rotary evaporator, and the white precipitate was removed by
filtration. Column chromatography on a 100 mm × 30 mm, 25-40
µm, LiChroprep RP-18 column (water/methanol gradient) at 1.5 mL/
min gave pure acetic acid [2-(4-methyl-2-oxocyclohexyl)propyl] ester
as a colorless liquid: yield, 134 mg (63% based on p-menthane-3,9-
diol); elemental composition (HRMS) C12H20O3, found (212.1381),
calcd (212.1412); MS (EI), m/z (relative intensity) 41 (27), 43 (63), 55
(27), 67 (17), 69 (29), 70 (18), 84 (15), 95 (14), 97 (20), 109 (15), 112
S-transferase during phase I (13) and phase II (10) metabolism
of pulegone.
From these data it seems that the metabolism of pulegone is
very well established, even though the most recent publication
(10) could not confirm some of the data published earlier,
basically, that menthofuran is a major metabolite. However, to
our knowledge there has never been a study focusing on the
metabolism of pulegone in humans in general. Therefore, both
enantiomers of pulegone were metabolized in MICA experi-
ments in this study. The aims were, first, to clarify whether the
metabolites found in animals can be also detected at ingestion
levels of pulegone that are comparable to a diet, second, to
clarify if there are other metabolites present that might have
been overlooked because of the early focus on menthofuran,
and, finally, to compare the data with the most recent publication
(10), which has shed some doubt on the findings of the earlier
publications.
1
(100), 113 (19), 137 (16), 212 (2); H NMR data (for the prevailing
diastereoisomer) (CDCl3) CH3-CH-CH2O (0.99, d, J ) 6.7 Hz, 3H),
CH3-CH-CH2-CO (1.02, d, J ) 6.3 Hz, 3H), CH3-CO (2.05, s,
3H), CH2-O (4.05, m, 2H), CH and CH2 of cyclohexane ring (2.37,
m, 1H; 2.27 m, 2H; 1.8-2.1, m, 4H; 1.3-1.5, m, 2H); 13C NMR
(CDCl3) CH3-CH-CH2O (15.25), CH3-CO (20.88), CH3-CH-CH2-
CO (22.21), CH2-CH2-CH-CO (28.92), CH2-CH2-CH-CO (33.96),
CH2-CO (50.72), CH2-O (66.87), CH3-CH-CH2CO (31.29), CH3-
CH-CH2O (35.30), CH-CO (52.29), CH3-COO (171.06), CH-CO
(211.00).
MATERIALS AND METHODS
2-(2-Hydroxy-1-methylethyl)-5-methylcyclohexanone (9-Hydroxy-p-
menthan-3-one, M1). The hydrolysis of pure 9-acetoxymenthone (10
mg) was performed using lithium methylate (5 µL of a 1.3 mmol/mL
solution in methanol) in methanol (1 mL), HCl (5 µL) in aqueous
tetrahydrofuran (1 mL; 50% w/w), Amberlyst (strongly acidic ion
exchanger; 100 mg) in aqueous tetrahydrofuran (1 mL; 50% w/w), and
potassium carbonate (50 mg) in aqueous methanol (1 mL; 50% w/w).
In all cases GC-MS of the reaction mixture showed three compounds
to be present in different proportions. It was impossible to purify any
of the three compounds.
Synthesis of 3-Hydroxy-3-methyl-6-(1-methylethyl)cyclohexanone
(1-Hydroxymenthone, M2). M2 was synthesized in a two-step
sequence according to a published procedure (17) with piperitone as
starting material. For NMR experiments 3-hydroxy-3-methyl-6-(1-
methylethyl)cyclohexanone was purified by column chromatography
on a 100 mm × 30 mm, 40 µm, Bakerbond Diol column (hexane/
diethyl ether gradient) at 1.5 mL/min.
Synthesis of E-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclo-
hexanone [10-Hydroxy-p-menth-4(8)-en-3-one, M4]. 10-Hydroxy-
p-menth-4(8)-en-3-one was synthesized in a four-step sequence starting
from isopulegol.
2-[1-(Chloromethyl)ethenyl]-5-methylcyclohexanol (10-Chloroiso-
pulegol). 10-Chloroisopulegol was synthesized according to a published
procedure (15) starting from isopulegol by chlorination with NaOCl/
CO2. It was used without further purification for the next step. The
yield was ∼90% on the basis of GC-MS. MS (EI), m/z (relative
intensity) 41 (48), 53 (30), 55 (53), 67 (60), 71 (100), 79 (42), 81 (83),
93 (56), 102 (48), 108 (44), 109 (48), 123 (40), 170 (15), 188 (M+; 8),
190 (M+; 3).
Chemicals. (S)-(-)-Pulegone, (R)-(+)-pulegone, and (+)-mentho-
furan were from Fluka (Buchs, Switzerland). Dess-Martin periodinane
was from Lancaster (Newgate, U.K.). â-Glucuronidase H2 from Helix
pomatia (100000 units/mL â-glucuronidase, 4500 units/mL sulfatase)
was from Sigma (Steinheim, Germany). A racemic mixture of p-men-
thane-3,9-diol was a generous gift from Dr. Stefano Serra (CNR,
Milano, Italy). Piperitone and mintlactone were generous gifts from
Dr. Akira Fujita (T. Hasegawa, Tokyo, Japan).
MICA Experiment. Human experimentation was performed as
previously described (14). The amount of (R)-(+)-pulegone applied
was 35 mg (0.24 mmol; ∼500 µg/kg of body mass), whereas the amount
of the less toxic (S)-(-)-pulegone was 70 mg (0.48 mmol; ∼1000 µg/
kg of body mass).
Isolation of Metabolites from Urine and Derivatization Proce-
dures. Isolation was performed as previously described (14). Briefly,
after enzymatic hydrolysis at pH 5.0 with glucuronidase and sulfatase
from H. pomatia and careful acidification to pH 2.0 with concentrated
HCl, the metabolites were subsequently extracted with diethyl ether
followed by a separation into neutral, phenolic, and acidic compounds.
The metabolites in the acidic fraction were trimethylsilylated with
BSTFA or ethylated with iodoethane (14).
Reference Compounds. Authentic standards were prepared accord-
ing to the procedures described below. All syntheses started from either
(S)-(-)-pulegone or (R)-(+)-pulegone, retaining the stereochemistry
at C-1 of the menthane skeleton.
Synthesis of 2-(2-Hydroxy-1-methylethyl)-5-methylcyclohexanone
(9-Hydroxy-p-menthan-3-one, M1). M1 was synthesized in a four-
step sequence starting from isopulegol.
2-Hydroxy-â,4-dimethylcyclohexaneethanol (p-Menthane-3,9-diol).
p-Menthane-3,9-diol was synthesized according to a published proce-
dure (15) starting from isopulegol by hydroborination and oxidation
with alkaline H2O2 solution. Independently, reference material was
obtained from CNR, Milano, Italy.
2-[1-(Chloromethyl)ethenyl]-5-methylcyclohexanone (10-Chloroiso-
pulegone). To a suspension of Dess-Martin periodinane (1.2 mmol)
in CH2Cl2 (10 mL) was added a solution of 10-chloroisopulegol (∼1
mmol) in CH2Cl2 (10 mL). The mixture was stirred for 1 h at room
temperature. The suspension was diluted with pentane (40 mL), the
precipitate removed by filtration, and the remaining solution concen-
trated to 5 mL using a rotary evaporator. The solvent was evaporated,
and the crude 10-chloroisopulegone was used for the next step without
further purification. MS (EI), m/z (relative intensity) 39 (63), 41 (58),
53 (45), 55 (27), 65 (29), 67 (65), 69 (39), 77 (26), 79 (43), 81 (37),
102 (30), 107 (68), 151 (100).
Acetic Acid [E-2-(2-Oxo-4-methylcyclohexylidene)propyl] Ester [10-
Acetoxy-p-menth-4(8)-en-3-one]. The crude 2-[1-(chloromethyl)ethen-
yl]-5-methylcyclohexanone of the previous step (∼1 mmol) was
dissolved in acetonitrile (4 mL). To this solution were added finely
ground anhydrous sodium acetate (10 mmol, 620 mg) and hexadecyl-
tributylphosphonium bromide (0.05 mmol, 25 mg). The suspension was
heated for 2 h in a closed vessel at 80 °C. The resulting mixture
consisted of approximately 30% menthofuran and 70% acetic acid [E-2-
(2-oxo-4-methylcyclohexylidene)propyl] ester. Purification was per-
formed by column chromatography on a 100 mm × 30 mm, 40 µm,
Acetic Acid [2-(2-Hydroxy-4-methylcyclohexyl)propyl] Ester (9-
Acetoxy-p-menthan-3-ol). To a solution of p-menthane-3,9-diol (1
mmol) in ethyl acetate (5 mL) were added acetic anhydride (1.0 mmol)
and 4-(dimethylamino)pyridine (0.1 mmol). The mixture was stirred
at room temperature until all of the diol was converted, as indicated
by GC-MS of the reaction mixture. The solution was evaporated, diluted
with CH2Cl2 (50 mL), and washed twice with water (5 mL), followed
by freshly prepared saturated NaHCO3 solution (5 mL). The solution
(mixture of two diastereoisomers) was dried over Na2SO4 and used
for the next step without further purification. MS (EI), m/z (relative
intensity) 43 (58), 55 (40), 67 (35), 68 (37), 69 (44), 81 (82), 95 (53),
97 (70), 107 (38), 112 (100), 121 (40), 136 (34), 214 (0.5). The MS
(EI) data were in excellent agreement with those previously published
(16).
Acetic Acid [2-(4-Methyl-2-oxocyclohexyl)propyl] Ester (9-Acetoxy-
p-menthan-3-one). To a suspension of Dess-Martin periodinane (1.2

Metabolism of Pulegone in Humans
J. Agric. Food Chem., Vol. 51, No. 22, 2003 6591
Bakerbond Diol column (40% tert-butyl methyl ether/60% hexane) at
1.5 mL/min. Yield, 90 mg (43% based on isopulegol); elemental
composition (HRMS) C12H18O3, found (210.1268), calcd (210.1256);
MS (EI), m/z (relative intensity) 39 (39), 41 (39), 43 (100), 53 (35), 55
(39), 65(14), 67 (32), 69 (12), 77 (14), 79 (37), 81 (20), 93 (15), 107
2 h at room temperature. The solution was washed with 5% citric acid
(2 × 5 mL), with saturated NaHCO3 solution (2 × 5 mL), and with
saturated NaCl solution. The remaining organic phase was dried over
Na2SO4 and concentrated to ∼10 mL by means of a rotary evaporator.
The product was sufficiently pure to be used for the next step without
further purification. An analytical sample for NMR experiments was
obtained by RP-HPLC on a 250 mm × 4.6 mm, 5 µm, Nucleosil C8
5-300 column at 1 mL/min with a water/methanol gradient. Elemental
composition (HRMS) C12H20O3, found (212.1429), calcd (212.1412);
MS (EI), m/z (relative intensity) 41 (47), 43 (95), 53 (20), 55 (37), 67
(49), 68 (27), 69 (39), 79 (33), 81 (82), 82 (25), 91 (27), 93 (64), 94
(20), 95 (55), 109 (100), 108 (27), 110 (21), 119 (21), 123 (93), 137
(44), 139 (21), 152 (M+ - CH3COOH, 80), 170 (2), 194 (M+ - H2O,
1
(20), 108 (38), 139 (29), 149 (13), 150 (32), 151 (33), 168 (16); H
NMR (CDCl3) CH3-CH (1.02, d, J ) 6.7 Hz, 3H), CH3-Cd (1.91,
m, 3H), CH3-CO (2.09, s, 3H), CH2-CH2-Cd (1.38, m, 1H; 1.90,
m, 1H), CH2-CO (2.05, m, 1H; 2.54, m, 1H), CH2-CH2-Cd (2.28,
m, 1H; 2.82, m, 1H), CH2-O (4.57, d, J ) 12.7 Hz, 1H; 4,68, d, J )
12.7 Hz, 1H), CH3-CH (2.05, m, 1H); 13C NMR (CDCl3) CH3-Cd
(17.84), CH3-CO (20.82), CH3-CH (21.69), CH2-CH2-Cd (28.72),
CH2-CH2-Cd (33.19), CH2-CO (51.39), CH2-O (64.36), CH3-CH
(32.32), CdC-CO (134.98), CdC-CO (136.72), CH3-COO (170.87),
CH2-CO (205.06).
1
5); H NMR (CDCl3) CH3-CH (1.03, d, J ) 6.7 Hz, 3H), CH3-Cd
(1.73, s, 3H), CH3-CO (2.05, s, 3H), CH2-CH2-Cd (1.25, m, 1H;
1.65, m, 1H), CH2-CO (1.44, m, 1H; 1.92, m, 1H), CH2-CH2-Cd
(1.96, m, 1H; 2.50, m, 1H), CH2-O (4.82, d, J ) 11.8 Hz, 1H; 4.95,
d, J ) 11.8 Hz, 1H), CH3-CH (2.05, m, 1H); CH-OH (4.58, m, 1H);
CH-OH (3.2, broad s, 1H); 13C NMR (CDCl3) CH3-Cd (17.22),
CH3-CO (21.15), CH3-CH (21.49), CH2-CH2-Cd (25.67), CH2-
CH2-Cd (33.21), CH2-CHOH (42.06), CH2-O (65.46), CH3-CH
(28.89), CH2-CHOH (70.48), CdC-CHOH (141.07), CdC-CHOH
(122.61), CH3-COO (171.98).
E-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclohexanone [10-hy-
droxy-p-menth-4(8)-en-3-one]. A solution of 10-acetoxy-p-menth-4(8)-
en-3-one (1 mmol) in methanol (3.5 mL) was mixed with NaOH
solution (3 mL; 0.1 mol/L). After exactly 5 min, hydrolysis was
terminated by the addition of KH2PO4 (0.3 mmol) dissolved in water
(10 mL). The mixture was immediately extracted with diethyl ether (2
× 30 mL). The combined organic phases were washed with water (1
× 1 mL) and brine (1 × 1 mL) and dried over Na2SO4. At this time a
GC-MS run was performed indicating that the reaction mixture
consisted of 20% menthofuran and 80% 10-hydroxy-p-menth-4(8)-en-
3-one. The mixture was concentrated to ∼1 mL, diluted with pentane
(2 mL), and immediately purified by chromatography on a 100 mm ×
30 mm, 40 µm, Bakerbond Diol column using a gradient starting with
40% diethyl ether/60% pentane at 10 °C and 3 mL/min. The fractions
(2 × 20 mL) containing the target compound were combined, and
CDCl3 (0.75 mL) was added and concentrated by means of a rotary
evaporator to ∼0.5 mL. The solution was again diluted with CDCl3
(0.75 mL) and concentrated to 0.6 mL. The solution contained ∼99%
10-hydroxy-p-menth-4(8)-en-3-one and 1% menthofuran at this time
besides a considerable amount of ether. The yield was in the range of
70-80% but was not determined exactly because the compound
decomposed rapidly and seemed to decompose more quickly the more
concentrated the solution. Elemental composition (HRMS) C10H16O2,
found (168.1133), calcd (168.1150); MS (EI), m/z (relative intensity)
41 (8), 55 (13), 67 (18), 69 (12), 77 (12), 79 (37), 80 (13), 81 (11), 91
(13), 93 (13), 107 (19), 108 (100), 109 (14), 135 (11), 139 (11), 149
(22), 150 (50), 168 (17); 1H NMR (CDCl3) CH3-CH (1.01, d, J ) 6.2
Hz, 3H), CH3-Cd (1.98, s, 3H), CH2-CH2-C) (1.37, m, 1H; 1.90,
m, 1H), CH2-CO (2.06, m, 1H; 2.52, m, 1H), CH2-CH2-Cd (2.25,
m, 1H; 2.81, m, 1H), CH2-O (4.13, d, J ) 12.8 Hz, 1H; 4.21, d, J )
12.8 Hz, 1H), CH3-CH (2.03, m, 1H); 13C NMR (CDCl3) CH3-Cd
(17.53), CH3-CH (21.55), CH2-CH2-Cd (28.03), CH2-CH2-Cd
(32.99), CH2-CO (51.10), CH2-O (62.62), CH3-CH (31.96), CdC-
CO (133.76), CdC-CO (141.07), CH2-CO (205.10).
Acetic Acid [Z-2-(2-Oxo-4-methylcyclohexylidene)propyl] Ester [9-Ac-
etoxy-p-menth-4(8)-en-3-one]. To Dess-Martin periodinane (1 mmol)
suspended in CH2Cl2 (50 mL) was added a solution of 9-acetoxy-p-
menth-4(8)-en-3-ol (∼0.9 mmol) in ethyl acetate (9 mL). The mixture
was stirred overnight at room temperature. It was taken to dryness,
and the white precipitate was extracted with pentane/diethyl ether [50:
50 (v/v); 4 × 40 mL]. The organic phase was concentrated and purified
by RP-HPLC on a 250 mm × 4.6 mm, 5 µm, Nucleosil C8 5-300
column at 1 mL/min with 25% methanol/75% water. Yield, 48% (101
mg) based on starting material mintlactone; elemental composition
(HRMS) C12H18O3, found (210.1265), calcd (210.1256); MS (EI), m/z
(relative intensity) 39 (11), 41 (17), 43 (50), 53 (12), 55 (16), 67 (13),
69 (12), 79 (30), 81 (14), 91 (13), 93 (15), 95 (12), 97 (14), 107 (18),
108 (100), 109 (19), 111 (12), 124 (10), 139 (90), 140 (10), 150 (M+
- CH3COOH; 65), 151 (29), 168 (38), 210 (M+; 1); 1H NMR (CDCl3)
CH3-CH (1.02, d, J ) 6.7 Hz, 3H), CH3-Cd (1.77, s, 3H), CH3-
CO (2.06, s, 3H), CH2-CH2-Cd (1.38, m, 1H; 1.90, m, 1H), CH2-
CO (2.06, m, 1H; 2.54, m, 1H), CH2-CH2-Cd (2.30, m, 1H; 2.74,
m, 1H), CH2-O (4.76, d, J ) 11.8 Hz, 1H; 4.93, d, J ) 11.8 Hz, 1H),
CH3-CH (2.00, m, 1H); 13C NMR (CDCl3) CH3-Cd (15.86), CH3-
CO (20.77), CH3-CH (21.53), CH2-CH2-Cd (28.70), CH2-CH2-
Cd (33.52), CH2-CO (50.56), CH2-O (65.55), CH3-CH (31.63), Cd
C-CO (136.05), CdC-CO (138.10), CH3-COO (170.66), CH2-CO
(203.59).
Z-2-(2-hydroxy-1-methylethylidene)-5-methyl-cyclohexanone; (9-hy-
droxy-p-menth-4(8)-en-3-one). Hydrolysis of 9-acetoxy-p-menth-4(8)-
en-3-one with either acid or weak base always led to menthofuran as
the single product. No trace of 9-hydroxy-p-menth-4(8)-en-3-one was
detectable by either HPLC or GC-MS.
Synthesis of Z-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclo-
hexanone [9-Hydroxy-p-menth-4(8)-en-3-one]. 9-Hydroxy-p-menth-
4(8)-en-3-one was synthesized in a four-step sequence starting from
mintlactone.
Synthesis of r,r,4-Trimethyl-1-cyclohexenemethanol (3-p-Men-
then-8-ol, M6). M6 was synthesized in a two-step sequence according
to a published procedure (18) starting from pulegone. For NMR
experiments, 4-trimethyl-1-cyclohexenemethanol was purified by chro-
matography on a 100 mm × 30 mm, 40 µm, Bakerbond Diol column
at 1.5 mL/min using a hexane/diethyl ether gradient. Yield, 78% based
on starting material pulegone; elemental composition (HRMS) C10H18O,
found (154.1341), calcd (154.1358); MS (EI), m/z (relative intensity)
43 (60), 59 (26), 79 (30), 81 (29), 93 (27), 95 (23), 107 (18), 121 (43),
Z-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclohexanol [p-Menth-
4(8)-ene-3,9-diol]. To a stirred suspension of LiAlH4 (1 mmol, 38 mg)
in dry diethyl ether (5 mL) was added dropwise a solution of
mintlactone (1 mmol, 166 mg) in dry diethyl ether (3 mL). After 15
min of stirring, the excess of LiAlH4 was destroyed by dropwise
addition of methyl formate in diethyl ether. The reaction mixture was
evaporated, and the remainder was treated with a solution of KH2PO4
(4 mmol, 544 mg) in water (10 mL), resulting in a pH of ∼7. The diol
was extracted with ethyl acetate (3 × 30 mL). The combined extracts
were washed with saturated NaCl solution (1 × 3 mL), and the organic
phase was dried over Na2SO4. The solution was used without further
purification for the next step. No effort was made to purify the diol
because of its poor chromatographic properties.
1
136 (17), 139 (100), 154 (15); H NMR (CDCl3) CH3-CH (0.95, d,
3H), CH3-CH (1.20, m, 1H), (CH3)2-C-OH (1.30, s, 3H and 1.31,
s, 3H), OH, CH2,CH2,CH2 (1.5-1.8 and 2.0-2.2, m, 7H), -CHdC
(5.70, m, 1H); 13C NMR (CDCl3) CH3-CH (21.67), CH3-CH (28.22),
(CH3)2-C-OH (28.89), CH2, CH2, CH2 (24.44; 31.34; 33.70), C-OH
(72.85), -CdCH (143.48), -CHdC (118.43).
Acetic Acid [Z-2-(2-Hydroxy-4-methylcyclohexylidene)propyl] Ester
[9-Acetoxy-p-menth-4(8)-en-3-ol]. To the crude solution of the p-menth-
4(8)-ene-3,9-diol in ethyl acetate (∼80 mL) were added acetic anhydride
(1.2 mmol) and 4-(dimethylamino)pyridine (0.1 mmol). The reaction
was monitored by GC-MS with the esterification being finished after
NMR Spectroscopy. NMR spectra (1H, 13C, 135° DEPT) were
recorded on a Bruker AM 360 instrument in CDCl3 (unless otherwise
stated) with TMS as internal standard (δ ) 0 ppm). Carbon multiplici-
ties of all compounds were determined by DEPT. Two-dimensional

6592 J. Agric. Food Chem., Vol. 51, No. 22, 2003
Engel
Figure 2. Mass spectrum of metabolite M1.
Figure 3. Synthesis of 9-acetoxymenthone (4), a precursor of M1.
experiments (COSY, HMBC, and HMQC) were recorded on a Bruker
AMX 400 instrument in CDCl3 (unless otherwise stated).
metabolites have a higher molecular weight because in most
cases oxygen is attached to the molecule during phase I reaction.
M1 was also present in untreated urine; however, its amount
differed considerably and ranged from extremely high to barely
detectable. Most probably M1 is bound to glucuronic acid or
sulfuric acid via a very labile bond, which might be hydrolyzed
partly even under the pH conditions of urine itself.
High-Resolution Gas Chromatography (HRGC))Mass Spec-
trometry (MS). HRGC was performed using a 30 m × 0.32 mm i.d.
DB-5 capillary column (J&W Scientific, Folsom, CA) in a gas
chromatograph type 8000 (ThermoQuest, Egelsbach, Germany). Sample
injection was performed using the split technique at a temperature of
230 °C and a split ratio of 1:25. The volumes injected ranged from 1
to 20 µL. Temperature during HRGC was held at 35 °C for 1 min and
raised to a final temperature of 230 °C at a rate of 6 °C/min. Mass
spectra were recorded on an MD 800 quadrupole mass spectrometer
(ThermoQuest) in the electron ionization (EI) mode at 70 eV.
HRGC)High-Resolution Mass Spectrometry (HR-MS). HRGC
was performed on a 30 m × 0.25 mm i.d. DB-5 capillary column (J&W
Scientific) in a gas chromatograph type 5890 (Hewlett-Packard,
Heilbronn, Germany). Sample injection was performed using the cold
on-column technique. The volumes injected ranged from 0.2 to 0.5
µL. Temperature during HRGC was held at 35 °C for 1 min and raised
to 230 °C at a rate of 6 °C/min. Mass spectra were recorded with a
Finnigan MAT 95S (Finnigan, Bremen, Germany) mass spectrometer
in the EI mode at 70 eV, in the chemical ionization (CI) mode at 115
eV with isobutane as reagent gas, and in high-resolution (HR) mode
using perfluorokerosene as internal standard.
From HRGC-HRMS the exact mass was consistent with the
elemental composition C10H16O. The molecular weight was also
confirmed by HRGC-MS-CI with isobutane as reagent gas.
Therefore, the very intense base peak with m/z 137 is formed
by removal of a methyl radical from the molecular ion with
m/z 152. On the basis of those facts it is most probable that the
methyl group is in an allylic position to a double bond present
in the molecule. In addition to the cyclic C6 ring system of
pulegone, it is most likely that there is a second ring structure
present in the molecule, therefore forming a partly hydrogenated
benzofuran skeleton. From a combination of all assumptions
the structure of M1 was proposed to be either 2,3,4,5,6,7-
hexahydro-3,6-dimethylbenzofuran (8,9-dihydromenthofuran, 5c
in Figure 3) or 2,3,3a,4,5,6-hexahydro-3,6-dimethylbenzofuran
(5b in Figure 3). Both compounds have the benzofuran ring
system and from both the removal of only one methyl group
during HRGC-MS, either from the cyclohexene ring or from
the dihydrofuran ring, is very probable.
To confirm the proposed structure for M1, the major
metabolite of (S)-(-)-pulegone, its precursor, 9-acetoxymen-
thone, was synthesized by the sequence shown in Figure 3.
The synthesis started from isopulegol (1), which was hydrobori-
nated and oxidized to the diol (2). The unhindered hydroxy
group was then esterified with acetic acid anhydride, forming
3, and the remaining hydroxy group converted to a carbonyl
function using Dess-Martin reagent, furnishing 9-acetoxy-
menthone (4). The hydrolysis of pure 9-acetoxymenthone was
performed using either lithium methylate in methanol, hydro-
chloric acid in aqueous tetrahydrofuran, a strongly acidic ion
exchanger in aqueous tetrahydrofuran, or potassium carbonate
in aqueous methanol. In all cases GC-MS of the reaction mixture
showed all three compounds (5a, 5b, and 5c) to be present in
different proportions. As an example, the GC-MS of the reaction
mixture dissolved in chloroform and methanol after hydrolysis
with lithium methylate is shown with mass spectra of the
corresponding peaks (Figure 4).
RESULTS AND DISCUSSION
The urine samples of six volunteers (three males and three
females) were collected for 24 h before (control) and after (test)
ingestion of 1000 µg of (S)-(-)- or 500 µg of (R)-(+)-pulegone/
kg of body weight. Because it can be assumed that most of the
metabolites of pulegone have been condensed with either
sulfuric acid or glucuronic acid during phase II reactions, the
samples were treated with glucuronidase and sulfatase to liberate
the apolar metabolites presumably formed in phase I reactions.
After extraction with diethyl ether and separation into acidic,
phenolic, and neutral compounds, each fraction was analyzed
by HRGC-MS either before or after trimethylsilylation or
ethylation. For analysis of unconjugated metabolites the urines
were extracted directly with diethyl ether after the pH had been
adjusted to either 2 or 7. All metabolites where detected in the
neutral fraction. Concentrated samples of test and control urine
were also tested by HPLC. However, no metabolites could be
detected in terms of new peaks present in the test sample (data
not shown). It would be desirable to analyze the conjugated
metabolites directly; however, only an HPLC method in
combination with radioactive 14C labeling would be reasonable,
but also inconsistent with experimentation on humans.
Metabolite M1. Surprisingly, the mass spectrum (Figure 2)
of the major metabolite of (S)-(-)-pulegone (based on GC-MS
area) was not identical with menthofuran as expected from the
results of most of the available published literature. It is
noteworthy that the molecular weight of the major metabolite
appeared to be the same as that of pulegone itself. Usually,
5a elutes last from the GC column and must therefore be the
alcohol 6-(2-hydroxy-1-methylethyl)-3-methylcyclohexanone (9-
hydroxymenthone). Because of the intense McLafferty rear-
rangement ion at m/z 112 the hemiketal tautomer can probably
be excluded to be present during gas chromatographic condi-
tions. 9-Hydroxymenthone is the major detectable product, in

Metabolism of Pulegone in Humans
J. Agric. Food Chem., Vol. 51, No. 22, 2003 6593
Figure 4. GC-MS of the hydrolysate of IV dissolved in chloroform (black trace) and methanol (gray trace).
the case when the methanol initially used is removed under
vacuum and replaced with chloroform. On the basis of GC-MS
analysis the methanolic solution consists mainly of peak 5c.
Discrimination between 5c and 5b is easily possible by
interpretation of the MS data. By retro-Diels-Alder rearrange-
ment from 5c an ion at m/z 110 is formed by loss of propene,
whereas from 5b the ion at m/z 124 is formed by loss of ethene
followed by loss of the furan methyl group, forming the ion at
m/z 109.
However, after enzymatic hydrolysis of the urine samples
only the structure 5c was found and never 5a or 5b. Because
M1 was not always found in the urines without treatment with
glucuronidase and sulfatase, it has to be concluded that the
original metabolite is most probably 5a or its cyclic hemiacetal.
Only these have a hydroxy group that can be linked to
glucuronic acid. The identified M1 might be formed spontane-
ously during enzymatic hydrolysis or during workup of the
enzymatically hydrolyzed urine. Complete loss of 5a and
formation of M1 during GC can be excluded, because chro-
matography of the synthetic mixture has proven that the
equilibrium is much more dependent on the solvent than on
temperature or injection technique.
It is noteworthy that it was not possible to purify any of the
three compounds and, therefore, NMR data for only acetylated
M1 were obtained. This is most probably because traces of water
present in any chromatographic system give rise to an equilib-
Figure 5. Polymorphism of M1 [only the 8-(S)-isomer is shown; numbering
refers to menthane skeleton].
rium between all possible structures (Figure 5). Additionally,
in more concentrated solutions as used for NMR the compound
will obviously form dimers to some extent by self-addition to
the enolether function, producing an even more complex
spectrum of compounds. The equilibrium might also explain
why this major metabolite has never been described in earlier
literature in cases when enzymatic hydrolysis was used to free
the primary metabolites from their conjugates. M1 is quite polar,
showing a behavior comparable to that of uroterpenolone (14),
a metabolite of carvone in humans, not being very well extracted
from aqueous solutions.
However, M1 was reported as being detected in a very recent
publication (10) and found to be conjugated to glucuronic acid.

6594 J. Agric. Food Chem., Vol. 51, No. 22, 2003
Engel
Figure 8. Synthesis of 3-hydroxy-3-methyl-6-(1-methylethyl)cyclohexanone.
at m/z 43 can be also explained by the open-chain form, which
contains two acetyl groups. Additionally, the hydroxy group is
attached in a â-position to the carbonyl group, from which
removal of water can be explained best. Finally, the position
of the hydroxy group at a tertiary carbon atom is well in line
with the expected behavior of phase I enzymes metabolizing a
saturated hydrocarbon.
Figure 6. Mass spectrum of metabolite M2.
To confirm the structural proposal, M2 was synthesized
according to a published (17) procedure (Figure 8) starting from
piperitone (11) via piperitone epoxide (12). The synthesized
sample showed identical retention times and mass spectrometric
decomposition and, therefore, M2 was successfully identified
as 3-hydroxy-3-methyl-6-(1-methylethyl)cyclohexanone (1-hy-
droxymenthan-3-one). This is the first identification of M2 as
a metabolite of pulegone in humans.
Figure 7. Possible structures of M2.
In conjugated form M1 cannot take part in the equilibrium
(Figure 5). It is noteworthy that only 1H NMR data are available
for the M1-glucuronic acid conjugate (10) and, therefore, it is
still not definite whether the open-chain form or the hemiketal
form is the one to be conjugated during phase II metabolism.
Metabolite M1 was also found as metabolite of the more toxic
(R)-(+)-pulegone; however, its amount was apparently much
lower compared to the amount found after ingestion of (S)-(-
)-pulegone. Because the hepatotoxicity of pulegone is linked,
at least in part, to menthofuran formation during metabolism,
the formation of 9-hydroxymenthone can be considered as a
true detoxification. Menthofuran formation should not be
possible from this metabolite. Therefore, the increased ability
of 9-hydroxymenthone formation from (S)-(-)-pulegone in the
body might very well explain the lower toxicity of the latter.
Metabolite M2. For metabolite 2 (M2) a mass spectrum was
obtained from the neutral fraction after enzymatic hydrolysis
(Figure 6). From HRGC-HRMS the exact mass was consistent
with the elemental composition C10H18O2. The molecular weight
was also confirmed by HRGC-MS in the CI mode. The
formation of m/z 152, that is, the loss of water, indicated that
there is at least one hydroxy group present in the molecule.
From the intense fragmentation pattern a hydrogenated benzo-
furan ring structure as in M1 could be excluded, and therefore
only a cyclohexanone ring system should be present. From the
quite intense fragment at m/z 128, which is most probably
formed from m/z 170 via McLafferty rearrangement by removal
of C3H6 from the isopropyl side chain, it was concluded that
the hydroxy group has to be attached to the ring system. The
following structures (Figure 7) were therefore considered as
theoretically possible structures of metabolite M2.
Metabolite M3. Metabolite M3 was identified as the well-
known compound menthol [3-methyl-6-(1-methylethyl)cyclo-
hexanol] by its mass spectrum and cochromatography with an
authentic standard. A total of three diastereoisomers of menthol
were found in the urine of test persons metabolizing (S)-(-)-
pulegone, but only one was found in those metabolizing (R)-
(+)-pulegone. The amount of menthol was apparently much
higher in the case of (S)-(-)-pulegone compared to (R)-(+)-
pulegone, again indicating that differences in toxicity are most
probably closely related to the ability of to reduce pulegone
rather than to metabolize it by oxidation. To our knowledge
this is the first identification of menthol as a metabolite of
pulegone.
Metabolite M4. Metabolite M4 was only a minor metabolite
of (S)-(-)-pulegone, but it was one of the major metabolites of
(R)-(+)-pulegone. For M4 a mass spectrum was obtained from
the neutral fraction showing a high degree of similarity
compared to the mass spectrum of menthofuran (Figure 9).
However, M4 was detected after enzymatic hydrolysis only,
again indicating conjugation to either glucuronic acid or sulfuric
acid.
From HRGC-HRMS the exact mass of the molecular ion at
m/z 168 was determined to be consistent with the elemental
composition C10H16O2. The molecular weight was also con-
firmed by HRGC-MS in the CI mode. The formation of the
ion at m/z 150, that is, the loss of water, indicated that there is
one hydroxy group present in the molecule. Because of the
similar mass spectra of menthofuran and M4, the latter was
considered to be a potential precursor for menthofuran earlier
identified as a metabolite of pulegone. Consequently, the only
useful possible structures for M4 were E- or Z-2-(2-hydroxy-
1-methylethylidene)-5-methylcyclohexanone. However, a review
of the literature revealed that neither compound has apparently
ever been synthesized or isolated and fully characterized.
Apparently, only a mass spectrum of Z-2-(2-hydroxy-1-meth-
ylethylidene)-5-methylcyclohexanone was published [available
as Supporting Information (19)] and without verification of the
proposed Z-configuration. Therefore, for both compounds
different synthetic methods were elaborated, which should allow
the unequivocal discrimination between both forms.
Structures 9 and 10 were considered to be unlikely as the
position of the oxygen is unusual because tertiary carbon atoms
are normally strongly favored over secondary carbon atoms by
phase I oxidizing enzymes. Structures 7 and 8, where the
hydroxy group is in an R-position to the carbonyl group, would
not explain the intense fragment at m/z 110, because the removal
of water is not favored from this position. Structure 6 seemed
ideal to be to explain the observed mass spectrum. It is the only
structure easily generating the fragment at m/z 58 after retro-
aldol reaction followed by McLafferty rearrangement under
mass spectrometric conditions. The intense formation of the ion
Z-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclohex-
anone (9-hydroxypulegone) was synthesized from mintlactone

Metabolism of Pulegone in Humans
J. Agric. Food Chem., Vol. 51, No. 22, 2003 6595
Figure 11. Synthesis of E-2-(2-hydroxy-1-methylethylidene)-5-methylcy-
clohexanone (10-hydroxypulegone, 21).
ethylidene)-5-methylcyclohexanone (10-acetoxypulegone, 21)
and menthofuran (5). This reaction consists of two steps, the
first being the isomerization of the double bond and intermediate
formation of 10-chloropulegone and 9-chloropulegone, depend-
ing on the conformation of the isopropenyl side chain. However,
9-chloropulegone is not stable under these conditions and
immediately cyclizes, forming menthofuran. In 10-chloropule-
gone cyclization is impossible because of the Z-configuration
of the double bond, and the chlorine atom is displaced in an
SN2 reaction by the unsolvated acetate anion. In the last step
the ester bond of 21 is hydrolyzed with dilute sodium hydroxide
solution in aqueous methanol. Menthofuran is stable under the
chosen hydrolysis conditions and can be removed during
purification of the product.
Figure 9. Comparison of mass spectra of menthofuran and metabolite
M4.
Surprisingly, in the first experiment after hydrolysis with
dilute K2CO3 solution in aqueous methanol for 1 h the yield of
10-hydroxypulegone was very low compared to that of the
byproduct, menthofuran, and it was not possible to obtain any
NMR data. Nevertheless, the synthesized sample showed
identical retention times and mass spectrometric fragmentation
during GC-MS and, therefore, M4 was successfully identified
as E-2-(2-hydroxy-1-methylethylidene)-5-methylcyclohexanone
(10-hydroxypulegone). This is the first identification of M4 as
a metabolite of pulegone in humans in vivo.
To find out more about the reactivity of 10-hydroxypulegone,
its stable precursor 10-acetoxypulegone was purified by column
chromatography and subjected to hydrolysis alone without
menthofuran being present. Surprisingly, depending on the
hydrolysis conditions, again ∼30% to as much as 100%
menthofuran was formed from 10-acetoxypulegone. The optimal
hydrolysis conditions to obtain 10-hydroxypulegone were treat-
ment of the precursor with a mixture of methanol/0.1 mol/L
NaOH (50:50, v/v) for a few minutes at room temperature. After
subsequent termination of hydrolysis and immediate extraction
of the target compound, column chromatography yielded pure
10-hydroxypulegone in sufficient amounts for NMR experi-
ments. It is noteworthy that even at -20 °C in the NMR solvent
CDCl3 under aprotic conditions, 10-hydroxypulegone decays
first very quickly and then slowly (Table 1), forming mentho-
furan as the sole product. Therefore, we cannot confirm the
stability observed previously (19).
The reaction of 10-hydroxypulegone forming menthofuran
is considerably faster in aqueous solution at room temperature
at any pH, leading to complete transformation of 10-hydroxy-
pulegone to menthofuran, usually within hours. Therefore, this
reaction has to be considered as a major source of menthofuran.
As a consequence, the amount of menthofuran detected in
metabolism experiments will be strongly dependent on the
Figure 10. Synthesis of Z-2-(2-hydroxy-1-methylethylidene)-5-methylcy-
clohexanone (9-hydroxypulegone, 17).
(13) in a four-step sequence (Figure 10). In the first step
mintlactone (13) was reduced, forming a diol (14) with the
required Z-configuration of the double bond. In the next step
the primary hydroxy group was protected by acetylation and
the resulting acetoxy alcohol (15) was oxidized to a direct
precursor (16) of the desired 9-hydroxypulegone (17). However,
during any kind of hydrolysis with different bases or under
acidic conditions no trace of 9-hydroxypulegone was ever
detected by either GC-MS or HPLC (chromatograms not
shown), and only menthofuran was formed from the acetylated
precursor (16) in quantitative yield within minutes. It has to be
concluded that 9-hydroxypulegone cannot exist and, therefore,
must be excluded as the structure for M4.
E-2-(2-Hydroxy-1-methylethylidene)-5-methylcyclohex-
anone (10-hydroxypulegone) was synthesized from isopulegol
(1) in a four-step procedure (Figure 11). In the first step
isopulegol was chlorinated and the resulting 10-chloroisopulegol
(18) oxidized using Dess-Martin periodinane, furnishing 10-
chloroisopulegone (19). Treatment of the latter with a suspension
of sodium acetate in acetonitrile in the presence of a phase
transfer catalyst gave a mixture of E-2-(2-acetoxy-1-methyl-

6596 J. Agric. Food Chem., Vol. 51, No. 22, 2003
Engel
Table 1. Decomposition of 10-Hydroxypulegone in CDCl3 at −20 °C
time (h)
10-hydroxypulegone (%)
methofuran (%)
0
4
24
48
120
1.2
4.9
7.7
8.4
8.7
98.9
95.1
92.3
91.6
91.3
workup method. Our experiments with highly purified 10-
hydroxypulegone (M4) support the opinion proposed very
recently (10) that menthofuran is most probably not a metabolite
of pulegone but an artifact resulting from workup. The precursor,
10-acetoxypulegone, is also not indefinitely stable and slowly
isomerizes, forming 9-acetoxypulegone. The driving force might
be explained with the orientation of the dipoles found in the
molecule. However, this conversion seemed to be apparently
slower (data not shown) compared to the decay of 10-
hydroxypulegone.
Minor Metabolites. There were some other metabolites
present in the test urines; however, all of their amounts were
very low compared to M1, M2, M3, and M4. Some of them
have been identified previously in rats. The following minor
metabolites were successfully identified by comparison with
authentic samples: piperitone (M5) and 4-trimethyl-1-cyclo-
hexene-1-methanol (M6, 3-p-menthen-8-ol). Piperitone might
also be formed as an artifact from M2 during enzymatic
hydrolysis. There were traces of other compounds present
only in test urines; however, due to their low amount it was
impossible to identify them by interpreting the mass spectro-
metric data. Because their amount was much lower compared
to the metabolites identified, they might also stem from
impurities in pulegone and therefore might not be metabolites
of the latter.
Menthofuran. After enzymatic hydrolysis, a mass spectrum
was obtained from test urines consistent with menthofuran. An
authentic sample of menthofuran showed identical retention
times and mass spectrometric fragmentation and, therefore, the
compound was identified as the latter. However, only trace
amounts of menthofuran were found in urines that were
extracted directly. As menthofuran itself has no hydroxy group
by which it could have been attached to either glucuronic acid
or sulfuric acid, it should have been present in the glucuronidase/
sulfatase-treated and in the untreated samples in equal amounts,
which was not the case.
The only explanation for this behavior is that menthofuran
is not formed in relevant amounts during metabolism but is
generated from one or more precursors during enzymatic
hydrolysis or workup as it was already shown for M4. However,
even after enzymatic hydrolysis, the amount of menthofuran is
much lower compared to M1 and also lower compared to M2.
Therefore, if menthofuran is really a major metabolite in
humans, it must have been further metabolized via epoxymen-
thofuran and hydroxymenthofuran, forming mintlactone (12).
Mintlactone could be only tentatively detected in the test urines
in this study, and only a noisy mass spectrum was obtained,
further excluding menthofuran as a major metabolite of pulegone
in humans. In recent literature the role of menthofuran as major
metabolite was questioned (1) because the amounts of mentho-
furan compared to pulegone found in blood were extremely low.
Additionally, menthofuran was also not found among the
metabolites of (R)-(+)-pulegone in F334 rats (10).
Figure 12. Proposed metabolism scheme for pulegone in humans.
(Figure 12) are proposed during metabolism of pulegone in
humans at ingestion-correlated nontoxic concentrations. In the
first step pulegone seems to be oxidized selectively in the 10-
position, forming 10-hydroxypulegone (M4). Alternatively, it
may be reduced to menthone, which was detected at trace levels
in all test urines (data not shown). It might be possible that
pulegone is also reduced at the carbonyl group first; however,
no trace of pulegol was found in the urine samples. Conse-
quently, pulegol is either reduced very efficiently to menthol
(M3) or rearranged under the conditions in the human body
forming R,R,4-trimethyl-1-cyclohexene-1-methanol (M6, 3-p-
menthen-8-ol), which was found as a minor metabolite. How-
ever, a more reasonable pathway leading to M3 is the reduction
of the carbonyl group in menthone. Metabolite M2 is a typical
menthone metabolite and therefore should be formed by
oxidation of the latter at C5.
In the literature allylomerization of pulegol has been reported
(18) to take place at a nearly physiologic pH of 6.5 under certain
conditions, leading to R,R,4-trimethyl-1-cyclohexene-1-methanol
(M6). However, the formation of M6 from pulegol as an artifact
during enzymatic hydrolysis with sulfatase and glucuronidase
also cannot be totally excluded. Some authors recently claimed
identification of pulegol as a metabolite of (S)-(-)-pulegone
(6). However, their published mass spectrum of pulegol was
totally different from the one we have obtained for synthetic
pulegol prepared according to a known method (18). In fact,
the mass spectra of pulegol and of M6 are very similar and,
therefore, likely to be confused, especially in cases when only
GC-MS has been used as an identification method without the
backup of synthesis.
The major metabolite M1 is formed from the primary
oxidation product M4 by reduction of the exocyclic double
bond. The polymorphism of M1 has already been discussed. It
remains unclear whether or to what extent M4 can be a precursor
for menthofuran in humans in vivo or if menthofuran has to be
considered solely as an artifact during workup. In a recent study
a nonvolatile metabolite in F344 rats was characterized (10),
which would also be a suitable precursor explaining the observed
formation of menthofuran during workup. However, at least at
Metabolism of Pulegone in Humans. On the basis of the
metabolites identified in this study, the following pathways

Metabolism of Pulegone in Humans
J. Agric. Food Chem., Vol. 51, No. 22, 2003 6597
(7) Madyastha, K. M.; Paul Raj, C. Studies on the metabolism of a
monoterpene ketone, R-(+)-pulegonesa hepatotoxin in rat:
Isolation and characterization of new metabolites. Xenobiotica
1993, 23, 509-518.
(8) Madyastha, K. M.; Paul Raj, C. Biotransformations of R-(+)-
pulegone and menthofuran in vitro: Chemical basis for toxicity.
Biochem. Biophys. Res. Commun. 1990, 173, 1086-1092.
(9) Madyastha, P.; Moorthy, B.; Vaidyanathan, C. S.; Madyastha,
K. M. In vivo and in vitro destruction of rat liver cytochrome
P-450 by a monoterpene ketone, pulegone. Biochem. Biophys.
Res. Commun. 1985, 128, 921-927.
(10) Chen, L.-J.; Lebetkin, E. H.; Burka, L. T. Metabolism of (R)-
(+)-pulegone in F344 rats. Drug Metab. Dispos. 2001, 29,
1567-1577.
(11) Madyastha, K. M.; Moorthy, B. Pulegone mediated hepatotox-
icity: Evidence for covalent binding of R-(+)-[14C]pulegone to
microsomal proteins in vitro. Chem.-Biol. Interact. 1989, 72,
325-333.
(12) Khojasteh-Bakht, S. C.; Chen, W.; Koenigs, L. L.; Peter, R. M.;
Nelson, S. D. Metabolism of (R)-(+)-pulegone and (R)-(+)-
menthofuran by human liver cytochrome P-450s: Evidence for
formation of a furan epoxide. Drug Metab. Dispos. 1999, 27,
574-580.
(13) Khojasteh-Bakht, S. C.; Nelson, S. D.; Atkins, W. M. Glutathione
S-transferase catalyzes the isomerization of (R)-2-hydroxymen-
thofuran to mintlactones. Arch. Biochem. Biophys. 1999, 370,
59-65.
low doses of pulegone, menthofuran is not a relevant metabolite
in humans in either (S)-(-)- or (R)-(+)-pulegone.
It seems that for (R)-(+)-pulegone both reduction steps to
menthone and from M4 to M1 are highly stereoselective and
much less favored by the conformation of the terpene compared
to (S)-(-)-pulegone, where M4 is accumulated. In case M4
cannot be removed quickly enough by reduction, further
oxidation at the hydroxy group might take place, forming the
obviously very reactive pulegone-10-aldehyde. This compound
might also be responsible for the toxic effects caused prefer-
entially by (R)-(+)-pulegone. Therefore, the differences in
toxicity between (R)-(+)- and (S)-(-)-pulegone can be very well
explained by the stereospecifity of the unknown enzyme
responsible for reducing the exocyclic double bond.
We were not able to clarify the observations published earlier
(20) that p-cresol is a major metabolite of menthofuran and,
therefore, also responsible for the toxic effects associated with
pulegone ingestion. p-Cresol was always present in the urines
of the test persons in amounts >100 times more compared to
the major metabolite M1 in both test and control urines. Further
studies with 13C-labeled pulegone are underway to clarify this
issue.
ACKNOWLEDGMENT
(14) Engel, W. In vivo studies on the metabolism of the monoterpenes
S-(+)- and R-(-)-carvone in humans using the metabolism of
ingestion-correlated amounts (MICA) approach. J. Agric. Food
Chem. 2001, 49, 4069-4075.
(15) Kreiser, W.; Ko¨rner, F. Stereospecific synthesis of (-)-â-
turmerone and (-)-bisacurol. HelV. Chim. Acta 1999, 82, 1610-
1629.
The support of Sibylle Suckart (technical assistance), Michael
Granvogl, Michael Czerny, Anja Fischer, Monika Berger, and
Ines Otte (participants in the MICA experiment), and Christine
Schwarz (NMR), Ines Otte (MS), and Julius Scho¨nauer (MS)
is gratefully acknowledged.
LITERATURE CITED
(16) Serra, S.; Fuganti, C. Enzyme-mediated preparation of enantio-
merically pure p-menthan-3,9-diols and their use for the synthesis
of natural p-menthane lactones and ethers. HelV. Chim. Acta
2002, 85, 2489-2502.
(17) Osuka, A.; Taka-Oka, K.; Suzuki, H. Chemo- and regioselective
reduction of R,â-epoxy ketones to -hydroxy ketones by sodium
hydrogentelluride. Chem. Lett. 1984, 271-272.
(18) Eschinasi, E. H. Synthesis of cis-pulegol and its allylomerization
products, 3-p-menthen-8-ol and 3-p-menthen-8-yl ethers. J. Org.
Chem. 1970, 35, 2010-2012.
(19) McClanahan, R. H.; Huitric, A. C.; Pearson, P. G.; Desper, J.
C.; Nelson, S. D. Evidence for a cytochrome P-450 catalyzed
allylic rearrangement with double bond topomerization. J. Am.
Chem. Soc. 1988, 110, 1979-1981.
(1) Anderson, I. B.; Mullen, W. H.; Meeker, J. E.; Khojasteh-Bakht,
S. C,; Oishi, S.; Nelson, S. D.; Blanc, P. D. Pennyroyal
toxicity: Measurement of toxic metabolite levels in two cases
and review of the literature. Ann. Intern. Med. 1996, 124, 726-
734.
(2) Bakerink, J. A.; Gospe, S. M., Jr.; Dimand, R. J.; Eldridge, M.
W. Multiple organ failure after ingestion of pennyroyal oil from
herbal tea in two infants. Pediatrics 1996, 98, 944-947.
(3) European Commission-SCF (Scientific Committee on Food).
Opinion of the Scientific Committee on Food on pulegone and
menthofuran. Opinion expressed on July 2, 2002; http://euro-
pa.eu.int/comm/food/fs/sc/scf/out133_en.pdf.
(4) Gordon, W. P.; Huitric, A. C.; Seth, C. L.; McClanahan, R. H.;
Nelson, S. D. The metabolism of the abortifacient terpene, (R)-
(+)-pulegone, to a proximate toxin, menthofuran. Drug Metab.
Dispos. 1987, 15, 589-594.
(20) Madyastha, K. M.; Paul Raj, C. Evidence for the formation of
a known toxin, p-cresol from menthofuran. Biochem. Biophys.
Res. Commun. 1991, 177, 440-446.
(5) Moorthy, B.; Madyashtha, P.; Madyastha, K. M. Metabolism of
a monoterpene ketone, R-(+)-pulegonesa hepatotoxin in rat.
Xenobiotica 1989, 19, 217-224.
(6) Madyastha, K. M.; Gaikwad, N. W. Metabolic fate of S-(-)-
pulegone in rat. Xenobiotica 1998, 28, 723-734.
Received for review June 30, 2003. Revised manuscript received August
12, 2003. Accepted August 12, 2003. We thank the Deutsche Fors-
chungsanstalt fu1r Lebensmittelchemie for financial support.
JF034702U