An approach to the chemotaxonomic differentiation of two European Dog's mercury species: Mercurialis annua L. and M. perennis L.

Article abstract of DOI:10.1002/cbdv.201100341

Mercurialis annua and M. perennis are medicinal plants used in complementary medicine. In the present work, analytical methods to allow a chemotaxonomic differentiation of M. annua and M. perennis by means of chemical marker compounds were established. In addition to previously published compounds, the exclusive presence of pyridine-3-carbonitrile and nicotinamide in CH₂Cl₂ extracts obtained from the herbal parts of M. annua was demonstrated by GC/MS. Notably, pyridine-3-carbonitrile was identified for the first time as a natural product. Further chromatographic separation of the CH₂Cl₂ extracts via polyamide yielded a MeOH fraction exhibiting a broad spectrum of side-chain saturated n-alkylresorcinols. While the n-alkylresorcinol pattern was similar for both plant species, some specific differences were observed for particular n-alkylresorcinol homologs. Finally, the investigation of H₂O extracts by LC/MS/MS revealed the presence of depside constituents. Whereas, in M. perennis, a mixture of mercurialis acid (=(2R)-[(E)-caffeoyl]-2-oxoglutarate) and phaselic acid (=(E)-caffeoyl-2-malate) could be detected, in M. annua solely phaselic acid was found. By comparison with synthesized enantiomerically pure (2R)- and (2S)-phaselic acids, the configuration of the depside could be determined as (2S) in M. annua and as (2R) in M. perennis.

Full text of DOI:10.1002/cbdv.201100341

282
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
An Approach to the Chemotaxonomic Differentiation of Two European
Dogꢀs Mercury Species: Mercurialis annua L. and M. perennis L.
by Peter Lorenz*, Sarina Duckstein, Jꢁrgen Conrad, Matthias Knçdler, Ulrich Meyer, and
Florian C. Stintzing
WALA Heilmittel GmbH, Department of Research and Development, Dorfstrasse 1,
D-73087 Bad Boll/Eckwꢀlden
(phone: þ49-7164930-6661; fax: þ49-7164930-228; e-mail: peter.lorenz@wala.de)
Mercurialis annua and M. perennis are medicinal plants used in complementary medicine. In the
present work, analytical methods to allow a chemotaxonomic differentiation of M. annua and M.
perennis by means of chemical marker compounds were established. In addition to previously published
compounds, the exclusive presence of pyridine-3-carbonitrile and nicotinamide in CH2Cl2 extracts
obtained from the herbal parts of M. annua was demonstrated by GC/MS. Notably, pyridine-3-
carbonitrile was identified for the first time as a natural product. Further chromatographic separation of
the CH2Cl2 extracts via polyamide yielded a MeOH fraction exhibiting a broad spectrum of side-chain
saturated n-alkylresorcinols. While the n-alkylresorcinol pattern was similar for both plant species, some
specific differences were observed for particular n-alkylresorcinol homologs. Finally, the investigation of
H2O extracts by LC/MS/MS revealed the presence of depside constituents. Whereas, in M. perennis, a
mixture of mercurialis acid (¼(2R)-[(E)-caffeoyl]-2-oxoglutarate) and phaselic acid (¼(E)-caffeoyl-2-
malate) could be detected, in M. annua solely phaselic acid was found. By comparison with synthesized
enantiomerically pure (2R)- and (2S)-phaselic acids, the configuration of the depside could be
determined as (2S) in M. annua and as (2R) in M. perennis.
Introduction. – The Euphorbiaceae comprise 300 genera with almost 7500 species
[1][2]. The subfamily Acalyphoideae (457 species), which is represented by genera
such as Acalypha, Dalechampia, Mallotus, Mercurialis, Ricinus, and Trewia [2][3], is
attracting a distinct pharmaceutical interest recently. In the course of a natural-product
screening program, the chemical constituents of a particular representative of this plant
family, the perennial dogꢁs mercury (Mercurialis perennis L.), was lately investigated in
depth. A broad spectrum of lipophilic and hydrophilic constituents of M. perennis have
been identified, e.g., alkaloids, terpenes, sterols, low molecular weight phenolics [4], n-
alkylresorcinols [5], depsides, as well as a great variety of flavonol glycosides [6].
Another widespread European species is annual dogꢁs mercury (M. annua L.).
Both M. perennis and M. annua have been used since ancient time in the European
herbal medicine for the treatment of various medical indications like difficult healing
wounds, sores and ulcers, burns and suppurations. Herbal parts of both plants are
recommended as a cure in constipations, as well as for the treatment of anorexia and
gynaecological disorders [7]. Although Mercurialis is no longer used in allopathy, the
aqueous fermented or hydroalcoholic preparations obtained from the herbal parts
continue to play a role in homeopathy and anthroposophic medicine [8]. However, a
clear differentiation of both Mercurialis species, particularly with regard to phyto-
ꢂ 2012 Verlag Helvetica Chimica Acta AG, Zꢃrich

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
283
chemical and pharmacological properties, has not been carried out. Also chemo-
taxonomic differences of both Mercurialis species are not reported in the literature.
Therefore, a close investigation on the phytochemical composition of M. annua was
performed to compare its metabolic compound pattern with that of M. perennis.
While M. perennis is a typical forest plant, primarily found in shady beech woods on
neutral or alkaline limestone soil (karst areas) [9][10], M. annua grows next to open
agricultural fields or on wasteland rich in fertilizers with a high tolerance to nitrates and
phosphates [7]. The flowering period of M. perennis is early spring (March–April),
while late summer (August–September) is typical for M. annua. It was hypothesized
that there are characteristic chemical-marker compounds in herbal parts for species
differentiation. For that reason, plant material for phytochemical investigations was
each collected during blossoming. Different types of extraction procedures were
chosen to evaluate the complete spectrum of both lipophilic and hydrophilic natural
constituents by using chromatographic and spectroscopic analyses.
Results and Discussion. – Lipophilic Volatile Constituents. Investigation of CH2Cl2
extracts from herbal parts of M. annua and M. perennis by GC/MS revealed several
compounds such as alkaloids, monoterpenes, and aromatic alcohols. Assignments were
performed according to the procedure described in [4]. In this manner, signals of seven
known constituents, i.e., benzyl alcohol (2), 2-phenylethyl alcohol (3), 6,6-dimethylbi-
cyclo[3.1.1]heptan-2-one (5), cis-myrtanal (6), cis-myrtanol (7) hermidine (9), and
hermidine quinone (10), were identified in extracts from both species (see Fig. 1, a and
b). Three further signals, however, could not be readily assigned. Mass spectra at tR 6.7,
9.1, and 16.0 min exhibited molecular-ion [Mþ ] peaks at m/z of 104, 121, and 122,
respectively. By comparison of the mass spectra with the NIST database [11] and using
commercially available reference compounds, these signals were identified as those
corresponding to pyridine-3-carbonitrile (1), N-ethylaniline (4), and nicotinamide (8).
Notably, all three constituents were detected for the first time in Mercurialis. While
compound 4 was found in both species, 1 and 8 were exclusively identified in M. annua
(Fig. 1, a). By GC/MS, the amounts of 1 and 8 were determined as 21.24 (ꢀ2.89) and
117.84 (ꢀ14.51) mg/kg, respectively, in the fresh plant material. To the best of our
knowledge, 1 has not been previously reported as a natural compound. However, a
closely related compound, 4-methoxypyridine-3-carbonitrile (11; see Fig. 2), the MeO
derivative of 1, has been earlier identified as a constituent of the tropical tree Hernandia
nyphaeifolia (Hernandiaceae family) [12]. Findings of the pyridine alkaloids 1 and 8 in
M. annua compelled us to more detailed chemotaxonomic and plant physiological
considerations. Nitrile-containing piperidine alkaloids have been described from
different tribes of the Euphorbiaceae family [3]. However, the Acalyphoideae family
exhibits a certain type of cyanogenesis and forms biogenetically relevant 1-methyl-oxo-
pyridine-3-carbonitriles, which in most cases are not cyanogenic [13]. For instance, ricinine
(12), ricinidine (13), and nudiflorine (14; see Fig. 2) were found in seeds of the castor
bean plant Ricinus communis and Trewia nudiflora [14][15]. Mallorepine (15) has been
identified in Mallotus repandus [16]. From herbal parts of Acalypha indica, cyanopiperi-
done alkaloid glucosides such as acalyphine (16) have been isolated [17][18]. The
presence of compound 1 in the same plant family is now established as another
structural homolog. From the literature, it is known that nicotinic acid (17) and nicotin-

284
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
285
Fig. 2. Known structurally relevant pyridine-3-carbonitrile and dihydro-oxo-pyridine-3-carbonitrile
alkaloids, formerly identified in members of the Acalyphoideae family (except 11 in the Hernandiaceae).
4-Methoxypyridine-3-carbonitrile (11), ricinine (¼1,2-dihydro-4-methoxy-1-methyl-2-oxopyridine-3-
carbonitrile; 12), ricinidine (¼1,2-dihydro-1-methyl-2-oxopyridine-3-carbonitrile; 13), nudiflorine
(¼1,6-dihydro-1-methyl-2-oxopyridine-6-carbonitrile; 14), mallorepine (¼1,4-dihydro-1-methyl-4-ox-
opyridine-3-carbonitrile; 15), acalyphine (¼1,2-dihydro-2-hydroxy-4-methoxy-1-methyl-6-oxopyridine-
3-carbonitrile 3-O-glucoside; 16).
amide (8) are the precursors for the biosynthesis of piperidine-carbonitrile alkaloids
(e.g., ricinine (12)), indicating a close association of these compounds with the pyridine
nucleotide (NAD) cycle [19][20]. Since the exact biosynthetic pathway is unkown, the
simultaneous presence of the two structurally relevant alkaloids 1 and 8 in M. annua led
to the assumption of a nicotinamide dehydratase/nitrile hydratase (nitrilase) enzyme
system in the plant. The first enzyme converts carboxylic acid amides into the
corresponding nitriles by dehydratation, and the second works in the opposite direction
(see Scheme 1). Nitrile hydratase has previously only been detected in bacteria and
fungi [21–23]. Its presence in higher plants would be a novelty. The exclusive presence
of pyridine-3-carbonitrile (1) in M. annua and earlier reports on the mammalian
toxicity after Mercurialis plant ingestion [24][25] open up the question if not hermidine
(9) and its derivatives but rather 1 would be the toxic principle of the plant. Since M.
perennis is typically found near shady woods, M. annua is omnipresent in many areas.
From a chemo-ecological point of view, the occurence of 1 in M. annua is, therefore,
plausible as a defense strategy against herbivores, while M. perennis assures its exis-
Scheme 1. Presumed Pyridine Alkaloid Biosynthesis Cycle in M. annua
a) –H2O, Nicotinamide dehydratase. b) þH2O, Nitrile hydratase. c) þ2 H2O (ꢁ NH3), Nitrilase.
d) þH2O, ( ꢁ NH3), Amidase; NADþ ¼nicotinamide adenin dinucleotide.

286
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
tence by a perennial rhizome system. However, further studies are warranted to investi-
gate the chemo-ecological role, biosynthetic pathways, and the toxicity of this alkaloid.
n-Alkylresorcinol Fingerprint. n-Alkylresorcinols were recently ascertained as
novel lipophilic constituents of M. perennis and the Euphorbiaceae family [5]. In the
present work, the chromatographic separation of the CH2Cl2 extracts of M. annua and
M. perennis via polyamide yielded alkylresorcinol enriched fractions (MeOH), which
were analyzed by GC/MS (mass scan on m/z 124): numerous side-chain saturated n-
alkylresorcinols (side-chain lengths C15 :0–C23 :0) were detected, with C19 :0 as the
main constituent (see Fig. 3, a and b). The identification of these structures was based
on published mass spectral data [5] and by use of a reference compound (3-
nonadecylresorcinol, C19 :0). While the n-alkylresorcinols were similar for both
Mercurialis species, characteristic differences were found for their specific distribution
(Fig. 3, c). Whereas, in M. annua, the shorter side-chain length homologs were
dominant (C15 :0, C17:0, C19 :0, and C21:0), a tendency for longer-chain homologs
(C19 :0, C21:0, and C23 :0) was observed for M. perennis. Thus, by analysis of the n-
alkylresorcinol profile a clear assignment of the corresponding Mercurialis species
appears to be feasible.
Hydrophilic Constituents. To profile hydrophilic constituents, H2O extracts
obtained from M. annua and M. perennis were investigated by LC(DAD)/MS/MS.
Characteristic UV chromatograms are exemplarily shown in Fig. 4, a and b. While in M.
annua, exclusively one main constituent at tR 32.9 min was detectable, in the extract of
M. perennis an additional signal at tR 41.5 min was detected. The negative-ion mass
spectra of these constituents exhibited peaks at m/z 591 and 619, representing [2 Mꢁ
H]ꢁ clusters, and their fragmentation into the pseudomolecular ions with peaks at m/z
295 and 309, respectively (see Fig. 5, a and b). In both mass spectra, the indicative
fragment peak at m/z 179 [caffeic acid ꢁH]ꢁ was observed. Furthermore, the compound
at tR 32.9 min showed a [malic acid ꢁH]ꢁ fragmentation and the other at tR 41.5 min a
cleavage of a [2-hydroxyglutaric acid ꢁH]ꢁ moiety. By close inspection of the MS data
and UV absorption characteristics (see Fig. 5), the depside constituents phaselic acid
(2-O-(E)-caffeoyl malic acid; 18) and mercurialis acid ((ꢁ)-(2R)-2-[(E)-caffeoyloxy]-
glutaric acid; 19) were identified [6]. Whereas in M. annua only 18 was found, in M.
perennis both compounds 18 and 19 were present. In the extract of M. perennis, the
peak-area ratio of 18 to 19 was determined as 1.00 :1.82 (see Fig. 4, b). The structure
and configuration of the depside 19 was recently determined by NMR and chiral-
resolution methods [6]. However, the absolute configuration of 18 in M. annua and M.
perennis is still unresolved. Since the occurrence of both (2R)- and (2S)-18 enantiomers
has been reported from several higher plant species [26–29] as well as from lichens
[30], it was interesting to elucidate the configuration of 18 in both Mercurialis species.
Configuration Determination of 18. For configuration determination of 18,
constituents of the H2O extracts of M. annua and M. perennis were hydrolyzed by
treatment with aqueous 1n NaOH. The resulting malic acid (23) was extracted with
AcOEt, followed by methylation with BF3/MeOH (see Scheme 2, sequence c–d). The
dimethylmalate enantiomers 21 thus obtained were separated on a chiral GC phase. In
the hydrolysate of M. annua, an almost 1:1 mixture (2R)-21/(2S)-21 was found, while in
M. perennis mainly (2R)-21 was present, beside small amounts of (2S)-21 (see Fig. 6).
The hydrolysis experiment indicated that in M. perennis the (2R)-18 enantiomer is

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
287
Fig. 3. Ion chromatograms (GC/MS, mass scan on m/z 124) of n-alkylresorcinol-enriched fractions
(MeOH eluate from polyamide) from M. annua (a) and M. perennis (b). Peaks identified via previously
published MS data [5]. c) Side-chain length distribution of n-alkylresorcinol homologs (C15 :0–C23 :0)
determined as relative abundance of peak areas (three separate workups, each measured in triplicate). A
generalized structural formula of the n-alkylresorcinol backbone is also given.

288
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
289

290
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
likely, but the absolute configuration of 18 in M. annua still remained unkown. An
enzymatic analysis of the free (unesterified) malic acid (23) in M. annua and M.
perennis revealed a predominance of the (2S)-23 (l-malate) in comparison to the (2R)-
23 enantiomer (d-malate), however, without configurational information on the
respective depsides (data not shown). Hence, an enantioselective synthesis of (2R)-18
and (2S)-18 was performed to obtain reference material for a chiral HPLC resolution
experiment (see Scheme 2). For the new synthesis of enantiomerically pure (2R)-18
and (2S)-18, an established protection-group concept was applied, formerly used by
Sefkow for the synthesis of chlorogenic acids [31]. The reaction of (E)-3,4-
bis(acetyloxy)cinnamoyl chloride (20) with the appropriate 21 enantiomer yielded
the intermediate condensation product 22. Subsequently, the protection groups were
removed from 22 by treatment with a 1n HCl/THF mixture to yield the product (2R)-18
or (2S)-18, respectively (Scheme 2). Finally, a separation of the synthesized enantiomer
mixture (2R)-18/(2S)-18 was achieved on a protein-based Chiral-AGPꢄ HPLC column
under reversed-phase conditions (see Fig. 7, g). By investigation of the Mercurialis H2O
extracts via chiral HPLC and spiking the samples with both enantiomerically pure
stereoisomers of 18, the presence of (2S)-18 in M. annua and (2R)-18 in M. perennis
was established unequivocally (Fig. 7, a–f). The opposite configuration of 18 in both
Mercurialis species leads to the question for their biosynthesis. Sullivan and Zarnowski
have recently identified a gene in red clover (Trifolium pratense) encoding a
hydroxycinnamoyl transferase (HCT-2), capable of transferring caffeoyl moieties
from a coenzyme A (CoA)-thiolester to l-malic acid ((2S)-23) [29][32]. Even if the
presence of such a hydroxycinnamoyl transferase in Mercurialis is likely, the different
substrate specifity for (2S)-23 or (2R)-23 to form the depside enantiomers (2R)-18 or
(2S)-18, respectively, remains unclear and is an interesting task for further studies.
Scheme 2. Synthesis of (2R)-Phaselic Acid ((2R)-18). Analogously, the (2S)-18 was synthesized. For
stereochemical analysis, the hydrolysis of (2R)- or (2S)-18 afforded (2R)- or (2S)-21, respectively, after
methylation.
a) Pyridine, CH2Cl2, 4-(dimethylamino)pyridine (DMAP), reflux. The arbitrary atom numbering is
shown for the intermediate 22. b) 1n HCl, THF, reflux. c) 1. 1n NaOH; 2. 1n HCl. d) BF3/MeOH.

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
291
Fig. 6. Chiral GC separation of hydrolyzed H2O extracts from M. annua (a) and M. perennis (b) after
treatment with BF3/MeOH. c) Reference mixture of dimethyl (2R)- and (2S)-malate ((2R)- and (2S)-21,
resp., 1:1 (w/w)).
Conclusions. – By using different extraction procedures and analytical methods,
lipophilic and hydrophilic metabolic profiles of M. annua and M. perennis were
analyzed. Specific chemical markers such as the newly found pyridine alkaloids, the

292
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
Fig. 7. HPLC Chromatograms (330 nm) of H2O extracts from M. annua (a) and M. perennis (d) on a
Chiral-AGPꢄ phase. b) M. annua spiked with (2R)-phaselic acid ((2R)-18); c) M. annua spiked with (2S)-
18; e) M. perennis spiked with (2R)-18; f) M. perennis spiked with (2S)-18; g) reference mixture of (2RS)-
18.

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
293
distinguished n-alkylresorcinol fingerprints, as well as the differences in the depside
composition and their respective configurations provide useful tools for a doubtless
phytochemical differentiation of M. annua and M. perennis. Among others, these
findings could be applied to the authentication of fresh plant material and herbal
remedies obtained from both species, and provide a basis for further biochemical
investigations. The phytochemical difference found may also trigger studies on the
pharmacological potency of M. perennis and M. annua preparations.
Experimental Part
Reagents and Standards. Pyridine-3-carbonitrile, 4-(dimethylamino)pyridine (DMAP), dimethyl
(þ)-(R)- and (ꢁ)-(S)-malate were purchased from Aldrich (D-Steinheim). n-Octadecane, (ꢁ)-cis-
myrtanol and C18-reversed-phase (RP) silica gel 100 were obtained from Fluka (CH-Buchs). N-
Ethylaniline and the boron trifluorideꢁmethanol complex (20% BF3 in MeOH (w/w)) were purchased
from Merck (D-Darmstadt). Nicotinamide and Amberliteꢄ XAD-7 HP were from SigmaꢁAldrich (D-
Taufkirchen), 3-nonadecylresorcinol, C19:0 from ReseaChem (CH-Burgdorf), 2-phenylethyl alcohol
and polyamide from Carl Roth (D-Karlsruhe). All other chemicals were obtained in anal. grade from
VWR (D-Darmstadt). Hermidine and (ꢁ)-cis-myrtanal were prepared according to the procedure
described in [4].
Plant Material. Herbal parts of M. annua were collected in September 2009 and 2010 on field borders
at the northwest side of Magdeburg (Saxony-Anhalt, Germany). Herbal parts of M. perennis were
collected in April 2009 and April 2010 in a limestone area of the mountain forest above Bad Boll/
Eckwaelden (Baden-Wuerttemberg, Germany). The fresh plant material was mechanically cleaned and
stored at ꢁ808 until investigation. Identification of the plant material was performed by Prof. O. Spring,
and voucher specimens are deposited with the herbarium of the Department of Botany, Hohenheim
University, Germany (M. annua, vouchers: HOH-012314, HOH-012315; M. perennis, vouchers: HOH-
011281, HOH-011282, and HOH-011283).
CH2Cl2 Extraction. GC/MS Quantification of Pyridine-3-carbonitrile (1) and Nicotinamide (8).
Frozen herbal parts of M. annua or M. perennis (50 g) were extracted with CH2Cl2 (500 ml) by
Ultrathurraxꢄ treatment (2 min at 15,200 rpm, IKA-Werke, D-Staufen). To prevent oxidative alterations,
the slurry was bubbled with N2 (10 min) before and after Ultrathurraxꢄ treatment. After standing for 24 h
at þ48, the plant material was filtered over Celite through a Bꢀchner funnel. Then, the pulp was washed
once with CH2Cl2 (50 ml) and extracted for a second time in the same manner. Subsequently, from the
unified extracts the solvent was distilled off by rotary evaporation in vacuo (at 358). For qual. GC/MS
analysis of 1 and 8, the residual spissum extract was dissolved in AcOEt (to 10 ml; containing the internal
standard n-octadecane) and submitted to GC/MS. Amounts of 1 and 8 were analyzed by use of external
calibration curves (calibration range: 0.16–6.44 mg/ml for compound 1, R2 ¼0.9997; and 0.17–6.94 mg/
ml for 8, R2 ¼0.9994).
GC/MS Profile of n-Alkylresorcinols. The plant material (100 g) was extracted twice with CH2Cl2
according to the previous protocol. The n-alkylresorcinol containing extract was concentrated according
to a modified literature procedure [5]. In brief, the spissum extract (dissolved in 30 ml CH2Cl2) was
loaded on a polyamide column (11ꢂ2 cm; particle size, 0.05–0.16 mm) preconditioned with CH2Cl2
(50 ml). Thereafter, the column was washed with CH2Cl2 (100 ml, fraction disposed), followed by MeOH
(100 ml). From the MeOH eluate, the solvent was removed in vacuo. To assess the n-alkylresorcinol
profile, the residue was re-dissolved in MeOH (1.5 ml) and analyzed by GC/MS (mass scan on m/z 124).
H2O Extraction. Herbal parts (50 g) were homogenized in distilled H2O (63 ml) by Ultrathurraxꢄ
treatment (5 min at 15.000 rpm). Before and after homogenization, the material was flushed with N2 to
prevent air oxidation. After standing over night at þ48, the plant material was squeezed over a cloth
strainer, and the obtained slurry was filtered over Celite through a Bꢀchner funnel by vacuum suction.
The resulting filtrates (ca. 70 ml) were stored at ꢁ808 until further investigation.

294
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
Anal. and Spectral Analyses. RP-HPLC/ESI-MS/MS Analyses were carried out on an Agilent 1200
HPLC system (Agilent Technologies Inc., Palo Alto, USA), equipped with a binary pump, a micro-
vacuum degasser, an autosampler, a thermostatic column compartment, and a UV/VIS diode array
detector. A ReproSil-Pur C18-AQꢄ RP column (5-mm particle size, 250ꢂ4.0 mm i.d., Dr. Maisch GmbH,
D-Ammerbuch-Entringen) was used for chromatographic separation at 238 and a flow rate of 1.0 ml/min.
The UV detection of the depsides (see Fig. 4) was performed at 330 nm. The mobile phase consisted of
HCOOH/H2O 0.2 :99.8 (v/v); mobile phase A) and MeCN (100%; mobile phase B). Starting with 10% B
for 25 min, a linear gradient was followed to 23% B at 60 min, remaining isocratically for 2 min, then
increasing to 100% B at 65 min, keeping for 5 min, before re-equilibration to starting conditions. The
H2O extracts of M. annua or M. perennis were centrifuged at 19,064 g (5 min) before use. The injection
volume of each sample was 20 ml. The LC system was coupled to an HCTultra ion trap (Bruker Daltonik
GmbH, D-Bremen) with an ESI source operating in the negative-ion mode with the following
parameters: HV voltage, 4000 V; dry gas, N2, 12.00 l/min with a dry temp. set at 3658; nebulizer, 70 psi.
Full-scan mass spectra of the HPLC eluates were recorded during the chromatographic separation
yielding [MꢁH]ꢁ ions. To obtain further structural information, these ions were trapped and fragmented
to yield the precursor product patterns of the analytes. The mass range of recording was m/z 50–2000
with a compound stability and trap drive level of 100%. MSn Data were acquired in the auto MS/MS
mode. The instruments were controlled by an Agilent ChemStation (B.01.03) and EsquireControl
Software (V6.1).
GC/MS Analyses. GC/MS was performed with a PerkinꢁElmer Clarus 500 gas chromatograph with
split injection (split ratio, 30 :1, injection volume, 1.0 ml) coupled to a mass detector. The column used was
a Zebron ZB-5ms cap. column (60 mꢂ0.25 mm i. d.ꢂ0.25 mm film thickness, 5% phenyl polysiloxane
and 95% dimethyl polysiloxane coating; Phenomenex, Torrance, USA). He was the carrier gas at a flow
rate of 1 ml/min. The injector used was a PSS (programmed-temp. split/splitless injector; temp., 2508).
The temp. program for the column oven was 100 to 3208 at 48/min with a final hold time of 30 min. The
mass spectrometer was run in the electron ionization (EI) mode (70 eV).
NMR Spectroscopy. The NMR spectra were recorded in CDCl3 at 500 (1H) and 125 MHz (13C) with a
Varian Unity Inova NMR spectrometer (D-Darmstadt). Chemical shifts are reported in d [ppm]
referenced to the residual solvent signal of CHCl3 (1H: d 7.26; 13C: d 77.00 ppm). 13C-NMR Signal
assignment of the novel compound 22 was based on 2D-heteronuclear NMR experiments (gHMBC and
gHSQC).
Synthesis of the Reference Compounds (2R)- and (2S)-Phaselic Acid ((2R)-18 and (2S)-18, resp.). 4-
[(1E)-3-Chloro-3-oxoprop-1-en-1-yl]benzene-1,2-diyl Diacetate (20). Compound 20 was synthesized in
95% yield according to a literature procedure [31]. M.p. 87–888 ([33]: 80–838). GC/MS Purity: GC/MS,
(70 eV) >99%; at tR 34.2 min: 284 (1), 282 (4, Mþ ), 247 (3, [MꢁCl]þ ), 242 (9), 240 (23, [MꢁCOMe]þ ),
205 (14, [MꢁCOMeCl]þ ), 200 (17), 198 (50, [Mꢁ2 COMe]þ ), 163 (100, [Mꢁ2 COMeCl]þ ), 134 (15).
The mass spectrum was essentially consistent with literature data [31].
Dimethyl (2R)-2-({(2E)-3-[3,4-Bis(acetyloxy)phenyl]prop-2-enoyl}oxy)butanedioate ((2R)-22). A
mixture of 20 (1.74 g, 6.16 mmol), dimethyl (þ)-(R)-malate ((2R)-21; 1.00 g, 6.17 mmol), DMAP
(0.068 g, 0.556 mmol), pyridine (40 ml), and CH2Cl2 (136 ml) was heated under reflux for 7 h.
Subsequently, the CH2Cl2/pyridine mixture was removed in vacuo by rotary evaporation, and the orange
residue was applied to vacuum liquid chromatography (VLC). In brief, TLC-grade SiO2 (75 g) was
preconditioned with hexane (100 ml). Then, a soln. of the crude material (dissolved in 20 ml of AcOEt)
was loaded on the column, which was again flushed with hexane (200 ml). By elution with a hexane/
AcOEt linear gradient (8 :2 to 3 :7 (v/v), several fractions were obtained, monitored by TLC (SiO2;
mobile phase: 100% AcOEt). Fractions 1 and 2 (hexane/AcOEt 4 :6 (v/v) containing the target
compound were combined, and the solvent was removed under reduced pressure. The residue obtained
(1.82 g) was washed with hexane (2ꢂ50 ml) by ultrasound treatment. Thus, remaining starting material
was removed with the supernatant. Further purification by repeated VLC afforded 1.36 g of pure (2R)-22
(colorless sirup; 54% of the theory). TLC (SiO2; 100% AcOEt): Rf 0.61. UV/VIS: 219 (4.21), 281 (4.36).
1H-NMR (CDCl3, 500 MHz): 7.66 (d, J ¼ 16.0, HꢁC(7’)); 7.39 (dd, J¼2.0, 8.6, HꢁC(6’)); 7.35 (d, J¼1.9,
HꢁC(2’)); 7.20 (d, J¼8.6, HꢁC(5’)); 6.42 (d, J¼15.7, HꢁC(8’)); 5.61–5.58 (m, HꢁC(2)); 3.77, 3.71 (2s, 2
MeO); 2.96–2.94 (m, CH2(3)); 2.28, 2.27 (2s, 2 Me of Ac). 13C-NMR (CDCl3, 125 MHz): 169.46, 169.25

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
295
(C(1), C(4)); 167.89, 167.79 (C(10’), C(12’)); 165.21 (C(9’)); 144.29 (C(7’)); 143.67 (C(4’)); 142.37 (C(3’));
132.83 (C(1’)); 126.44 (C(6’)); 123.87 (C(5’)); 122.83 (C(2’)); 117.82 (C(8’)); 68.31 (C(2)); 52.65, 52.11
(C(5), C(6)); 35.97 (C(3)); 20.53, 20.48 (C(11’), C(13’)). The 13C-NMR assignments were based on
gHSQC and gHMBC experiments. The arbitrary C-atom numbering is shown in Scheme 2. GC/MS at tR
49.4 min; (70 eV): 408 (6, Mþ ), 377 (4), 366 (16, [MꢁCOMe]þ ), 324 (65, [M–2 COMe]þ ), 247 (7), 205
(13), 162 (100), 145 (4), 134 (48), 113 (8). The enantiomer (2S)-22 was prepared via a similar procedure.
(2R)-22: [a]2D5 ¼–5.00 (c¼1.00, MeOH); (2S)-22: [a]2D5 ¼ þ4.93 (c¼1.01, MeOH).
(2R)-Phaselic Acid (¼(2R)-2-{[(2E)-3-(3,4-Dihydroxyphenyl)prop-2-enoyl]oxy}butanedioic Acid;
(2R)-18). In a mixture THF/1n HCl 1:3 (v/v, 75 ml), (2R)-22 (1.15 g, 2.816 mmol) was dissolved under
N2. Then, the mixture was stirred and heated under reflux for 6 h. Afterwards, the pH was adjusted to 3.0
by dropwise addition of 1n NaOH, and the soln. was saturated with NaCl and extracted with AcOEt (3ꢂ
100 ml). The combined AcOEt extracts were dried (Na2SO4), and the solvent was removed in vacuo by
rotary evaporation to yield 0.94 g of crude material. By VLC on a C18-RP-SiO2 column (100 g of TLC
grade 100 C18-RP SiO2, preconditioned subsequently with MeOH and MeOH/H2O, 10 : 90 (v/v)), the
product was purified by elution with a linear MeOH/H2O gradient from 10 :90 to 30 :70 (v/v). Fractions
were analyzed by TLC showing that Frs. 13–15 (MeOH/H2O 30 :70 (v/v)) contained the target product.
From the combined fractions, the solvent was distilled off in vacuo to yield pure (2R)-18 (0.41 g; 49% of
the theory). Pale yellow resin. TLC (SiO2; CHCl3/AcOH/H2O 10 :8.4 :1.6 (v/v/v)): Rf 0.32. HPLC Purity:
>99.9%, tR 35.7 min. [a]2D4 ¼ ꢁ16.48 (c¼1.64, MeOH); [a]2D3 ¼ ꢁ32.9 (c¼2.06, H2O); [30]: [a]D ¼ ꢁ37.7
(c¼2.18, H2O). UV/VIS: 219 (4.13), 245 (4.00), 331 (4.26)1). Compound (2S)-18 was prepared in 51.5%
yield in the same manner from the corresponding precursor. (2S)-18: [a]2D5 ¼ ꢁ15.96 (c¼1.75, MeOH);
[a]2D3 ¼ þ29.6 (c¼1.75, H2O); [26]: [a]D ¼ þ31.5 (c¼1.36, H2O)1).
Configuration Analysis of Malic Acids, (2R)-23 and (2S)-23, by Hydrolysis of the H2O Extract, and
GC Separation of (2R)-21 and (2S)-21 after Derivatization. The H2O extracts of M. annua and M.
perennis (35 ml each) were stirred for 24 h with 1n NaOH (10 ml). The resulting brown solns. were
acidified with 1n HCl to pH 1.2 and diluted with H2O (50 ml). The H2O layers were saturated with NaCl
and extracted with AcOEt (2ꢂ50 ml). Then, the combined AcOEt extracts were dried (Na2SO4), the
solvent was removed by vacuum rotovaporation, and the residues obtained were treated with 20% BF3 in
MeOH (w/v; 1 ml) for 45 min at 1058. Subsequently, the mixtures were transferred into sat. NaCl soln.
(5 ml) and extracted with hexane (1ꢂ5 ml). The hexane extracts were dried (Na2SO4) and evaporated to
a small volume (1 ml) by a vigorous stream of N2. From the soln. thus obtained, 1 ml was injected into a
GC-FID (PerkinꢁElmer, Autosystem XL), equipped with a chiral cap. column (IVADEX-7, 25 mꢂ
0.25 mm i. d.ꢂ0.25 mm film thickness, diethyl tert-butylsilyl b-cyclodextrine coating; IVA Analysentech-
nik, D-Meerbusch). He was used as carrier gas at a flow rate of 1.3 ml/min. The injector and detector
temps. (FID) were set to 2308. The temp. program for the column oven was: 50–1808 at 1.58/min with a
total run time of 87 min. For comparison dimethyl (þ)-(R)- and (ꢁ)-(S)-malate ((2R)-21 and (2S)-21,
resp.) were used as reference.
Analysis of Free (2R)-23 and (2S)-23 by Enzymatic Reaction. To remove polyphenols and colorants,
H2O extracts of M. annua and M. perennis (4 ml each) were filtered twice by vacuum suction over an
Amberliteꢄ XAD-7 HP column (2-cm diameter, 2.5-cm length; washed before with 100 ml H2O). The
obtained filtrates were analyzed by use of commercially available d- and l-malic acid assays (R-
Biopharm AG, D-Darmstadt). The amounts of d- or l-malic acids ((2R)- and (2S)-23, resp.) were
determined via increase of NADH (by absorbance at 340 nm) which is formed by enzymatic conversion
of malic acid, nicotinamide-adenine dinucteotide (NADþ ), and d- or l-malate dehydrogenase, resp.: d-
(l)-malic acidþNADþ !oxaloacetateþNADHþHþ [34].
Chiral HPLC Separation of (2R)-18 and (2S)-18 in the H2O Extracts. The configuration analysis of 18
was performed on a Dionex Ultimate 3000 HPLC (Dionex GmbH, D-Idstein). A Chiral-AGPꢄ column
(ChromTec-Ltd., 5-mm particle size, 150ꢂ2.0 mm-i. d., Dr. Maisch GmbH, D-Ammerbuch-Entringen)
was used for chromatographic separation at 208. The UV detection of the phaselic acid enantiomers was
performed at 220 and 330 nm (Fig. 7). Using NaH2PO4 in H2O (0.1m, pH 3.0, eluent A) and MeCN
1
)
The anal. data were in accordance with those in the literature [26][30]. See also Fig. 5, a, for LC/MS/
MS data.

296
CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
(100%; eluent B), the following gradient at a constant flow rate of 0.21 ml/min was applied: 0–20 min,
0% B; 21–23 min, 0–10% B; 23–28 min, 10% B; 29–30 min, 10–0% B; 31–36 min, 0% B. The software
used for HPLC system control and data processing was Chromeleon (V6.8; Dionex GmbH, D-Idstein).
The H2O extracts of M. perennis and M. annua (70 ml) were filtered by vacuum suction over SiO2 (for
TLC, layer: 1.5ꢂ3.5 cm), and the filtrate was centrifuged for 5 min at 14,000 rpm before use. The
injection volume of each sample was 5 ml. Compounds (2R)-18 and (2S)-18 (2.0 mg/l and 1.8 mg/l resp.)
were employed as reference.
The authors wish to acknowledge Prof. O. Spring (Department of Botany, Hohenheim University, D-
Stuttgart) for identification of the plant material.
REFERENCES
[1] D. Frohne, U. Jensen, ꢅSystematik des Pflanzenreichsꢁ, 4th edn., Gustav Fischer Verlag, Stuttgart,
Jena, New York, 1992, p. 220.
[2] W. Erhardt, E. Gçtz, N. Bçdecker, S. Seybold, ꢅDer große Zander–Enzyklopꢀdie der Pflanzenna-
menꢁ, Vol. 2, ꢅFamilie und Gattungenꢁ, Eugen Ulmer, Stuttgart (Hohenheim), 2008.
[3] R. Hegnauer, ꢅChemotaxonomie der Pflanzenꢁ, Birkhꢀuser Verlag, Basel, Boston, Berlin, 1989, Vol.
VIII, p. 453.
[4] P. Lorenz, M. Hradecky, M. Berger, J. Bertrams, U. Meyer, F. C. Stintzing, Phytochem. Anal. 2010,
21, 234.
[5] P. Lorenz, M. Knoedler, J. Bertrams, M. Berger, U. Meyer, F. C. Stintzing, Z. Naturforsch., C/J. Biosci.
2010, 65, 174.
[6] P. Lorenz, J. Conrad, J. Bertrams, M. Berger, S. Duckstein, U. Meyer, F. C. Stintzing, Phytochem.
Anal. 2012, 23, 60.
[7] G. Madaus, ꢅLehrbuch der biologischen Heilmittelꢁ, reprint of the edn. Leipzig 1938, Medimed,
Hildesheim, 1979, 2nd edn., p. 1892.
[8] Deutscher Apotheker Verlag, ꢅGerman Homoeopathic Pharmacopeia (HAB): Monographs E–Z,
Vol. 2ꢁ, Govi-Verlag, Stuttgart – Pharmazeutischer Verlag, Eschborn, 2000.
[9] R. Pott, J. Hꢃppe, ꢅSpezielle Geobotanikꢁ, Springer-Verlag, Berlin, Heidelberg, 2007, p. 197.
[10] R. G. Jefferson, J. Ecol. 2008, 96, 386.
[11] S. Stein, Y. Mirokhin, D. Tchekhovskoi, G. Mallard, ꢅThe NIST Mass Spectral Search Programꢁ,
NIST/EPA/NIH Spectral Library, Vers. 2.0, distributed by National Institute of Standard and
Technology (USA), 2008.
[12] K. Yakushijin, S. Sugiyama, Y. Mori, H. Murata, H. Furukawa, Phytochemistry 1980, 19, 161.
[13] J.-D. Kant, Ph.D. Thesis, Technical University Braunschweig, 1983.
[14] G. R. Waller, L. M. Henderson, J. Biol. Chem. 1961, 236, 1186.
[15] G. B. Fodor, B. Colasanti, ꢅThe Pyridine and Piperidine Alkaloids: Chemistry and Pharmacologyꢁ, in
ꢅAlkaloids: Chemical and Biological Perspectives, Vol. 3ꢁ, Ed. S. W. Pelletier, John Wiley & Sons,
New York, 1985.
[16] H. Hikino, M. Tamada, K.-Y. Yen, Planta Med. 1978, 33, 385.
[17] A. Nahrstedt, J.-D. Kant, V. Wray, Phytochemistry 1982, 21, 101.
[18] M. Hungeling, M. Lechtenberg, F. R. Fronczek, A. Nahrstedt, Phytochemistry 2009, 70, 270.
[19] R. D. Johnson, G. R. Waller, Phytochemistry 1974, 13, 1493.
[20] A. Matsui, Y. Yin, K. Yamanaka, M. Iwasaki, H. Ashihara, Physiol. Plant. 2007, 131, 191.
[21] T. Nagasawa, C. D. Mathew, J. Mauger, H. Yamada, Appl. Environ. Microbiol. 1988, 56, 1766.
[22] Y. Kato, S. Yoshida, S.-X. Xie, Y. Asano, J. Biosci. Bioeng. 2004, 97, 250.
ˇ
´
[23] O. Kaplan, V. Vejvoda, O. Plꢆhal, P. Pompach, D. Kavan, P. Bojarova, K. Bezouska, M. Mackova, M.
˚
ˇ
´
Cantarella, V. Jirku, V. Kren, L. Martꢆnkova, Appl. Microbiol. Biotechnol. 2006, 73, 567.
[24] R. Bismarck, W. Floehr, Dtsch. Tierꢁrztl. Wochenschr. 1974, 81, 433.
[25] P. J. Watson, Vet. Rec. 1998, 142, 116.
[26] R. Hahn, A. Nahrstedt, Planta Med. 1993, 59, 71; R. Hahn, A. Nahrstedt, Planta Med. 1993, 59, 189.

CHEMISTRY & BIODIVERSITY – Vol. 9 (2012)
297
[27] N. Didry, V. Seidel, L. Dubreuil, F. Tillequin, F. Bailleul, J. Ethnopharmacol. 1999, 67, 197.
[28] S. Sahpaz, N. Garbacki, M. Tits, F. Bailleul, J. Ethnopharmacol. 2002, 79, 389.
[29] M. L. Sullivan, R. Zarnowski, Planta 2010, 231, 319.
[30] H. Tazaki, M. Ito, M. Miyoshi, J. Kawabata, E. Fukushi, T. Fujita, M. Motouri, T. Furuki, K. Nabeta,
Biosci. Biotechnol. Biochem. 2002, 66, 255.
[31] M. Sefkow, Eur. J. Org. Chem. 2001, 1137.
[32] M. L. Sullivan, Plant Physiol. 2009, 150, 1866.
[33] R. Balint, G. Cooper, M. Staebell, P. Filner, J. Biol. Chem. 1987, 262, 11026.
[34] H. Mçllering, ꢅMethods of enzymatic analysisꢁ, 2nd edn., Ed. H. U. Bergmeyer, Verlag Chemie,
Weinheim/Academic Press, New York and London, 1974, Vol. 3, p. 1589.
Received October 20, 2011