Synthesis, in vitro reactivity, and identification of 6-phenyl-3-hexen-2- one in human urine after kava-kava (Piper methysticum) ingestion

Article abstract of DOI:10.1055/s-2005-837781

This study describes the synthesis of 6-phenyl-3-hexen-2-one, a proposed metabolite of kava-kava (kava, 'Awa, Yaqona, Piper methysticum Forst.), its reactivity with glutathione in vitro, and its isolation and identification, as its mercapturic acid adduct using LC/MS/MS, in the urine of two human subjects following their ingestion of kava. A possible metabolic pathway for the formation of this metabolite and its possible role in hepatotoxicity are also discussed.

Full text of DOI:10.1055/s-2005-837781

Synthesis, in vitro Reactivity, and Identification of
6-Phenyl-3-hexen-2-one in Human Urine after Kava-
Kava (Piper methysticum)Ingestion
Lihong Zou
Martha R. Harkey
Gary L. Henderson
Abstract
mation of this metabolite and its possible role in hepatotoxicity
are also discussed.
This study describes the synthesis of 6 -phenyl-3-hexen-2-one, a
proposed metabolite of kava-kava (kava, `Awa, Yaqona, Piper Key words
methysticum Forst.), its reactivity with glutathione in vitro, and Kava-kava ´ kava ´ metabolism ´ glutathione ´ hepatotoxicity ´ re-
its isolation and identification, as its mercapturic acid adduct active metabolites ´ 6-phenyl-3-hexen-2-one ´ Piper methysticum ´
using LC/MS/MS, in the urine of two human subjects following Piperaceae
their ingestion of kava. A possible metabolic pathway for the for-
Introduction
Materials and Methods
The formation of biologically reactive metabolites and subsequent Glutathione, N-acetylcysteine, and ammonium formate were
reaction with cellular proteins has been proposed as a possible purchased from Sigma Chemical Co. (St. Louis, MO). Analytical
mechanism for the extremely rare cases of hepatotoxicity thought standards of kavain (K), dihydrokavain (DK), methysticin (M), di-
to be associated with the consumption of herbal products contain- hydromethysticin (DM), yangonin (Y), and desmethoxyyangonin
ing kava-kava (kava, `Awa, Yaqona, Piper methysticum Forst.). Sev- (DY) were obtained from PhytoLab GmbH & Co. (Vestenbergs-
eral glutathione (GSH) reactive metabolites of kava-kava have greuth, Germany). The supplier provided a certificate of analysis
been identified in in vitro studies; however, as yet, none have for these compounds and their purity was verified in our labora-
been found in human urine [1], [2]. Koppel et al. identified 6-phe- tory by LC/MS/MS. Powdered, dry kava-kava root was obtained
nyl-3-hexen-2-one (6-PHO) as a human urinary metabolite of ka- from Kava Kauai, Kapaa, HI. Quantitation of the six major kava-
vain [3] but, because the urine samples were hydrolyzed prior to lactones present in the root material was accomplished by LC/
analysis, it is not clear whether this compound was excreted as MS/MS. 4-Phenyl-1-butene, CDCl3, DMSO-d6, and 1-triphenyl-
its free form or as a conjugate. We reasoned that because 6-PHO phosphoranylidene-2-propanone were purchased from Aldrich
is an a,b-unsaturated ketone, it is a ªsoftº electrophile and there- Chemical Co. (Milwaukee, WI). All other chemicals were obtain-
fore likely to be highly reactive. In this communication, we report ed from Fisher Scientific (Fair Lawn, NJ). Solid-phase, C18 extrac-
the synthesis of 6-PHO, demonstrate its reactivity in vitro with tion columns were obtained from Varian Sample Preparation
GSH, and its presence (as a mercapturic acid adduct) in the urine (Harbor City, CA). High purity (HPLC grade) solvents were used
142
of kava-kava drinkers.
in all analytical procedures.
Affiliation
Department of Medical Pharmacology and Toxicology, School of Medicine, University of California, Davis,
CA, USA
Correspondence
Gary L. Henderson, Ph. D. ´ Department of Medical Pharmacology and Toxicology ´ School of Medicine ´
University of California ´ Davis ´ California 95616 ´ USA ´ Phone: +1-530-752-8141 ´ Fax: +1-530-752-4256 ´
E-mail: glhenderson@ucdavis.edu
Received April 6, 2004 ´ Accepted August 28, 2004
Bibliography
Planta Med 2005; 71: 142±146 ´ ꢀ Georg Thieme Verlag KG Stuttgart ´ New York
DOI 10.1055/s-2005-837781
ISSN 0032-0943

Instrumental analysis
eluted with 2 mL of 50% ACN (v/v), then lyophilized to dryness
Metabolite identification and quantitation of the lactones in the and reconstituted with 1 mL of MeOH.
kava-kava beverage was accomplished using a ThermoFinnigan
LCQ mass spectrometer (ThermoFinnigan, San Jose, CA), oper- Proposed ESI + LC/MS/MS: MH+: 338, CID (20%) MH+: m/z = 164
ated in either the atmospheric pressure chemical ionization [H2SCH2CH(COOH)NHC = OCH3]+, m/z = 175 [C6H5-(CH2)2CH =
(APCI) or electrospray ionization (ESI) mode. The mass spectro- CHC = OCH3H]+, m/z = 122 [HSCH2CH(COOH)NH2]+. 1H-NMR
meter was coupled to the outlet of an Agilent HP1100 HPLC sys- (CDCl3): d = 1.9 (sb, 2H), 2.15 (s, 3H), 2.2 (s, 3H), 2.73 (m, 1H),
tem equipped with a C18 column (4.6 mm”250 mm, 5 mm) or C8 2.79 (t, 2H), 3.13 (m, 2H), 3.06 (m, 2H, cys-b,b¢), 4.8 (m, 1H, cys-
column (4.6 mm”150 mm, 3.5 mm) (Thermo Hypersil-Keystone, CHCOOH), 6.6 (d, 1H, cys-NH), 7.3 (m, 5H, phenyl ring). 13C-NMR
Bellefonte, PA) at ambient temperature (228C). NMR spectra (CDCl3): d = 22.8 (CH3), 30.6 (CH3), 32.3(cys-b), 32.9 (CH2), 38.2
were obtained using a Bruker Avance-500 MHz NMR spectrome- (CH2), 41.8 (CH2), 48.9 (CH), 52.8 (cys-a), 128.5 (CH”2, Ar), 128.4
ter and recorded at room temperature (228C) with Me4Si at 0 (CH, Ar), 128.3 (CH, Ar), 126.2 (CH, Ar), 141.1(C, Ar), 171.9 (CO
ppm as the reference.
amide), 172.5 (COOH), 207.8 (CO).
Synthesis of 6-PHO
In vitro reactivity of 6-PHO with GSH
6-PHO was synthesized according to the general method of Hon Non-enzymatic reactivity of 6-PHO with GSH under physiologic-
et al. [4]. Briefly, a stream of ozone was bubbled through a solu- al conditions was investigated by mixing 6-PHO (0.01 mM) with
tion of 5 mL CH2Cl2 and 4-phenyl-1-butene (1 mmol) at ±788C. GSH (10 mM) at 378C and pH 7.4, then immediately extracting
Ozone treatment was terminated when the solution assumed a and analyzing the mixture as described above. LC/MS/MS analy-
blue color, then a stream of N2 was used to remove excess ozone. sis of the extract revealed a peak with a protonated molecule of
The reaction mixture was allowed to warm to room temperature m/z = 482 and with a retention time and fragmentation pattern
(228C) then a mixture of Et3N (1.1 mmol) and Ph3 = CHCO2Me consistent with the synthesized GSH derivative. Collision-in-
(1.1 mmol) in 5 mL CH2Cl2 was added and the solution stirred duced dissociation of the protonated molecule (shown in Fig.1)
for 15 min. Proposed APCI + LC/MS/MS: MH+: 175. CID (25%) of generated product ions characteristic of a GSH adduct at m/z =
MH+: m/z = 157 [MH±H2O]+, m/z = 131 [C6H5CH2CH2CH = CH]+, 407, 353, and 308, which we propose are due to the loss of pyro-
m/z = 117 [C6H5CH2CH = CH]+, m/z = 105 [C6H5CH2CH2]+, m/ glycine (75 Da), pyroglutamic acid (129 Da) and 6-phenyl-3-hex-
z = 71 [CH2CH2C = OCH3]+. 1H-NMR (CD3Cl): d = 2.2 (s, 3H), en-2-one (174 Da), respectively [5].
2.58 (dd, 2H, J = 2.77 Hz), 2.8 (t, 2H, J = 7.13 Hz), 6.1 (d, 1H,
J = 15.7 Hz), 6.8 (m, 1H, J = 6.75 Hz), 7.15±7.3 (m, 5H, J = 7.56 In vivo formation of 6-PHO-GSH
Hz, J = 7.28 Hz).
In vivo reactivity of 6-PHO with GSH was demonstrated by iden-
tifying the corresponding N-acetylcysteine adduct of 6-PHO in
human urine. Urine samples were provided by self-identified
Synthesis of 6-PHO-GSH
The glutathione conjugate of 6-PHO was synthesized by adding regular drinkers of kava-kava (Caucasian, one male and one fe-
400 mmol 6-PHO in MeOH (5 mL) to GSH (500 mmol in 10 mL male) who had not consumed kava-kava for at least one week be-
0.5 M phosphate potassium buffer at pH 8.0) over a period of 30 fore consuming approximately 10 g of powdered, dry kava-kava
min. The solution was stirred for an additional 30 min, extracted root mixed in water. Each individual provided a urine sample col-
with CHCl3 to remove side products, the aqueous layer lyophi- lected immediate before they ingested kava-kava, a urine sample
lized to dryness, reconstituted in 0.2% AcOH (1 mL), then added obtained on their first voiding after kava-kava consumption, and
to a Bond Elutꢁ C18 solid phase extraction cartridge. The column a urine sample collected the following day. Urine samples were
was washed with 2 mL of 0.2% AcOH (v/v) then eluted with 2 mL frozen immediately after collection and arrived at our laboratory
143
of 30% ACN (v/v).
frozen. After thawing, aliquots (30 mL) were centrifuged at
4,000”g, the supernatant removed and acidified to pH 2.0 with
Proposed ESI + LC/MS/MS: MH+: 482, CID (20%) MH+: m/z = 407 AcOH, then extracted with EtOAc (60 mL”5). Combined organic
[MH±pyroglycine]+, m/z = 353 [MH±pyroglutamic acid]+, m/z =
extracts were dried under vacuum and reconstituted in 0.5 mL of
335 [MH±pyroglutamic acid±H2O]+, and m/z = 308 (GSH)+. 1H- methanol. Samples were analyzed in random order with the ana-
NMR (DMSO-d6): d = 1.8 (sb, 2H), 1.9 (m, 2H, glu-b,b¢), 2.1 (s, lyst ªblindedº as to the identity of the samples. Aliquots (10 mL)
3H, CH3), 2.3 (m, 2H, glu-g,g¢), 2.72 (m, 1H), 2.75 (d, 2H, cys-b,b¢), were injected onto a C18 column then chromatographed using a
2.79 (m, 2H), 3.1 (m, 2H), 3.3 (m, 2H, glu-a,a¢), 3.6 (sb, 2H, gly- solvent system consisting of A (10 mM of ammonium formate)
a,a¢), 4.4 (m, 1H, cys-a,a¢), 7.1±7.3 (m, 5H, phenyl ring), 8.3 (sb, and B (ACN:MeOH:H2O, 42:42:16). Mobile phase B was in-
1H, gly-NH), 8.5 (d, 1H, cys-NH).
creased linearly from 30±80% within 35 min at a flow rate 1
mL/min. Selected ions, formed at the ESI ion spray interface at a
voltage of 4.5 kV in a positive mode, were transmitted through
Synthesis of the N-acetylcysteine adduct 6-PHO
The N-acetylcysteine adduct of 6-PHO was synthesized by add- the heated capillary at 2508C with collision-induced dissociation
ing 0.1 mmol of 6-PHO in MeOH (2 mL) to 0.14 mmol of N-acet- at 20% collision energy and helium as the collision gas.
yl-L-cysteine along with 50 mL of 200 mM phosphate potassium
buffer (pH 8.0) as a catalyst. The reaction medium was stirred for Extracted standards, prepared by fortifying pre-dose urine sam-
1 h, the solvent removed by rotary vacuum, and the residue re- ples with 50 ng/mL (N = 2) and 300 ng/mL (N = 2) of the syn-
constituted in 1 mL of the mixture of ACN and 0.2% AcOH (v/v, thesized N-acetylcysteine adduct of 6-PHO, were extracted and
25%:75%). The solution was applied to a Bond Elutꢁ C18 solid analyzed as described above. The absolute recovery of the extrac-
phase extraction cartridge, washed with 2 mL of 0.2% AcOH, tion method was found to be 99  3% (50 ng/mL) and 92  4%
Zou L et al. Synthesis, in vitro¼ Planta Med 2005; 71: 142±146

(300 ng/mL). To estimate the amount of the N-acetylcysteine ad- tion samples. In addition, the abundance of ion m/z = 338 in the
duct of 6-PHO in human urine, a calibration curve was prepared immediate post-consumption sample was significantly higher
by fortifying pre-dose urine samples with 10, 50, 100, 250, and than the sample obtained the following day. Retention time of
1,000 ng/mL of the synthesized N-acetylcysteine adduct of 6- the m/z = 338 peak was consistent with that of a peak found in
PHO then extracting and analyzing the samples as described a pre-consumption urine fortified with the synthesized standard
above.
of the N-acetylcysteine adduct of 6-PHO. Further confirmation of
this structure was accomplished by fragmenting the protonated
A sample of the dried kava-kava root powder ingested was ana- molecule m/z = 338 (Fig. 3) which resulted in a fragmentation
lyzed by LC/MS/MS for the six major kavalactones using a proce- pattern identical to that of the extracted, synthesized standard
dure described previously [6]. Total kavalactone concentration of the N-acetylcysteine adduct of 6-PHO. Selected reaction-mon-
(Table 1) was 13%, which is in range of reported levels (3±20%) itoring (SRM) LC/MS/MS analysis using two characteristic mass
found in the dried root [7], but considerably lower than some transitions: m/z = 338 ® 164 and 338 ® 122 (Fig. 4) provided
herbal preparations standardized to contain 70% kavalactones additional evidence for the mercapturic acid adduct in urine. Co-
[8]. Ratios of individual lactones (DK > K > DM > M > DY > Y) incident responses were observed in each ion current chromato-
were generally consistent with those reported in the dried plant gram at a retention time of 14.8 min, corresponding to the reten-
[9] and similar to those used in phytomedicines.
tion time of the synthetic N-acetylcysteine adduct of 6-PHO.
Semi-quantitative analysis of the urine samples (no internal
standard was used) showed the N-acetylcysteine adduct of 6-
PHO present in low (approximately 0.6 mg/mL) concentrations
in the immediate post-consumption sample and too low to esti-
Results
The in vitro reactivity of 6-PHO with GSH under physiological mate in the urine sample obtained the next day. Similar results
conditions (378 C, pH 7.4) is demonstrated by data in Fig.1. A were obtained from the urine samples from both kava-kava
compound formed immediately which LC/MS/MS analysis re- users; however, data for only one subject is shown.
vealed to be identical to synthesized 6-PHO-GSH. Fig. 2 shows
single ion monitoring (SIM) chromatograms comparing pre- and
post-consumption urine samples obtained from one of the hu- Discussion
man subjects. A single peak for ion m/z = 338 (representing the
protonated molecule for the mercapturic acid adduct) was found Our findings suggest that 6-PHO is indeed a highly reactive me-
in both post-consumption samples, but not in the pre-consump- tabolite of kava-kava, which can be detected in human urine as a
Table 1 Concentrations of the six major kava lactones in the dried kava root consumed by the subjects
144
Kava Lactones
MW
Concentration
(mg/mg extract)
Concentration
(% extract)
Concentration
(mmol/mg extract)
Desmethoxyyangonin
Dihydrokavain
Dihydromethysticin
Kavain
228.26
232.26
276.26
231.26
274.26
258.26
2.67
75.73
17.17
29.82
4.93
0.27
7.57
0.01
0.33
0.06
0.13
0.02
0.0002
0.55
1.71
2.98
Methysticin
0.49
Yangonin
0.05
0.005
13.03
Total Kava Lactones
130.37
Fig. 1 Product ion mass spectrum of 6-phenyl-3-hexen-2-
one-SGH formed in vitro obtained by collision-induced disso-
ciation of the protonated molecule MH+ m/z = 482. Product
ions at m/z = 407, 353, and 308 are due to the loss of pyro-
glycine (75 Da), pyroglutamic acid (129 Da) and 6-phenyl-3-
hexen-2-one (174 Da), respectively. This spectrum is identi-
cal to that of the synthesized standard.
Zou L et al. Synthesis, in vitro¼ Planta Med 2005; 71: 142±146

6-PHO could be formed by scission of the pyrone ring of dihydro-
kavain followed by a series of reactions including decarboxyla-
tion, O-demethylation and oxidation. Kavain may also be a
source of 6-PHO because dihydrokavain (DK) has been found
after administration of kavain [3], [11], [12]. Another ªsoft elec-
trophileº, K7, might also be produced from kavain by this series
of reactions and is also likely to be highly reactive, but it was
not investigated in the present study.
There could be environmental sources of 6-PHO as well. Once ka-
vain or dihydrokavain are hydrolyzed, the resulting acids decar-
boxylate and oxidize rather readily. Rasmussen et al. [12] found
6-PHO as a urinary metabolite when dihydrokavaic acid (DK4)
was administered to rats (400 mg/kg, p.o.) but noted that 6-PHO
could be produced by the action of alkali on dihydrokavain. They
also observed that when DK4 was added to control urine, sub-
stantial amounts were converted to DK6. Thus, 6-PHO could oc-
cur as a component of the kava-kava plant, a side-product
formed during the harvesting or processing of the plant, or an ar-
tifact formed during the preparation of analytical samples. We
did not find 6-PHO in the kava-kava product consumed or in the
pre-consumption urine; therefore, it is most likely that 6-PHO
identified in this study was indeed formed in vivo. However, the
possibility that this material could be formed during the proces-
sing of kava-kava into an herbal product should be considered
when evaluating kava-associated heptotoxicity.
In summary, we have demonstrated the reactivity of the kava-
kava metabolite 6-PHO with GSH in vitro and identified its mer-
capturic acid derivative in human urine. It is possible that other
nucleophiles such as protein thiols or DNA bases could conjugate
with 6-PHO and result in toxicity by disrupting the activity of cri-
tical enzymes, alkylating DNA, or combining with hepatic pro-
teins to produce an immune-mediated toxic reaction. It should
be noted, however, that reactive metabolites do not invariably
produce toxicity. Reactive species occur relatively commonly in
the metabolic pathways of drugs and other xenobiotics and for-
Fig. 2 SIM (single ion monitoring) chromatograms obtained from a
human urine sample before and after ingesting kava. Chromatograms
show intensity of current for ion m/z = 338, the protonated molecule of
the N-acetylcysteine adduct of 6-PHO, and were obtained using ESI +LC/
MS.
145
mercapturic acid adduct. A proposed metabolic pathway for the eign proteins or haptens usually do not produce a significant im-
formation of 6-PHO is shown in Fig. 5. K1 and DK1 (hydroxyla- mune response without a stimulus (adjuvant) that upregulates
tion at C-12 of the aromatic ring) are reported to be the most co-stimulatory molecules (see review by Uetrecht [13]). Never-
abundant urinary metabolites (excreted as glucuronide and sul- theless, it could be useful to confirm the presence of this reactive
fate conjugates) following oral administration of kava-kava [10] metabolite in a larger group of kava-kava users. Of particular in-
or kavain to humans [3], [11] and rats [12]. K2 [3], [11], [12], terest would be determining whether this metabolite is present
DK2 [3], resulting from dehydration of the 5,6 position, and K3, in the urine of South Pacific Islanders, the traditional users of
resulting from O-demethylation of the lactone ring [12] have kava-kava who exhibit little or no toxicity to kava-kava, and
been identified as well as a number of ring-open metabolites in- who reportedly metabolize a wide variety of drugs to a signifi-
cluding K4 [3], K6 [3], K7 [3], [11], DK6 [12], and 6-PHO [3]. Thus, cantly lower degree than do Europeans [14].
Fig. 3 Product ion mass spectrum of an N-acetylcysteine
adduct of 6-PHO found in human urine obtained by collision
induced dissociation of the protonated molecule MH+ m/z =
338. The ion at m/z = 164 is from the protonated form of N-
acetylcysteine.
Zou L et al. Synthesis, in vitro¼ Planta Med 2005; 71: 142±146

References
1
Johnson BM, Qiu SX, Zhang S, Zhang F, Burdette JE, Yu L, et al. Identifi-
cation of novel electrophilic metabolites of Piper methysticum Forst.
(Kava). Chem Res Toxicol 2003; 16: 733±40
2 White KD, Hartman N, Strong J, Musser SM. Metabolism of kava kava
pyrones to glutathione reactive metabolites. 51st American Society of
Mass Spectrometry, July 8±12, 2003 Montreal, Canada:
3 Koppel C, Tenczer J. Mass spectral characterization of urinary metabo-
lites of d,L-kawain. J Chromatogr 1991; 562: 207±11
4 Hon Y, Lu L. An extremely efficient way to prepare conjugated carbonyl
compounds from terminal alkenes via the reactions of ozonides, tri-
ethylamine and stable phosphorus ylides. Tetrahedron Letters 1995;
51: 7937±42
5 Baillie TA, Davis MR. Mass spectrometry in the analysis of glutathione
conjugates. Biol Mass Spectrom 1993; 22: 319±25
6 Zou L, Henderson GL, Harkey MR, Sakai Y, Li AP. Effects of kava (kava-
kava, ºAwa, Yaqona, Piper methysticum) on c-DNA-expressed cyto-
chrome P450 enzymes and human cryopreserved hepatocytes. Phyto-
medicine 2004; 11: 285±94
Fig. 4 SRM (selected reaction monitoring) chromatograms from a ur-
ine sample obtained from a human volunteer after ingestion of kava.
Chromatograms show intensity of ion current for the collision-induced
dissociation of the protonated molecule m/z = 338 to m/z = 164 (up-
per trace) and m/z = 122 (lower trace). Retention time is consistent
with that of extracted synthesized N-acetylcysteine adduct of 6-PHO.
7 Lebot V, Levesque J. Genetic control of kavalactone chemotypes in
Piper methysticum cultivars. Phytochemistry 1996; 43: 397±403
8 Malsch U, Kieser M. Efficacy of kava-kava in the treatment of non-psy-
chotic anxiety, following pretreatment with benzodiazepines. Psycho-
pharmacology (Berl) 2001; 157: 277±83
9 Kubatova A, Miller DJ, Hawthorne SB. Comparison of subcritical water
and organic solvents for extracting kava lactones from kava root. J
Chromatogr 2001; 923: 187±94
Acknowledgements
10
Duffield AM, Jamieson DD, Lidgard RO, Duffield PH, Bourne DJ. Identi-
The authors wish to thank Drs. Yung-Son Hon, Institute of Chem-
istry, Taiwan and Krishnan P. Nambiar, University of California,
Davis for their valuable advice on the synthesis of 6-phenyl-3-
hexen-2-one. We also gratefully acknowledge the help of Jeffery
DeRopp of the UC Davis NMR facility and NSFgrant OSTI 97±
2441 for funding the NMR instrumentation.
fication of some human urinary metabolites of the intoxicating bever-
age kava. J Chromatogr 1989; 475: 273±81
Tarbah F, Mahler H, Kardel B, Weinmann W, Hafner D, Daldrup T. Ki-
11
netics of kavain and its metabolites after oral application. J Chroma-
togr B Analyt Technol Biomed Life Sci 2003; 789: 115±30
Rasmussen AK, Scheline RR, Solheim E, Hansel R. Metabolism of some
12
kava pyrones in the rat. Xenobiotica 1979; 9: 1±16
Uetrecht J. Screening for the potential of a drug candidate to cause
13
idiosyncratic drug reactions. Drug Discovery Today 2003; 8: 832±37
Kaneko A, Lum JK, Yaviong L, Takahashi N, Ishizaki T, Bertilsson L et al.
High and variable frequencies of CYP2C19 mutations: medical conse-
quences of poor drug metabolism in Vanuatu and other Pacific islands.
Pharmacogenetics 1999; 9: 581±90
The research described was supported by grant number R01
AT000636 from the National Center for Complementary and
Alternative Medicine (NCCAM) and its contents are solely the
responsibility of the authors and do not necessarily represent
the official views of the NCCAM or the National Institutes of
Health.
14
146
Fig. 5 Proposed metabolic pathway for ka-
vain (K) and dihydrokavain (DK) leading to
the formation of the reactive metabolite 6-
PHO. Numbers in parenthesis refer to refer-
ences in which the respective metabolite
was identified.
Zou L et al. Synthesis, in vitro¼ Planta Med 2005; 71: 142±146