Microsomal metabolism of dictamnine: Identification of metabolites and evaluation of their mutagenicity in Salmonella typhimurium

Article abstract of DOI:10.1093/mutage/14.2.181

Incubation of the furoquinoline alkaloid dictamnine with microsomal fractions from phenobarbital-induced rat liver resulted in a metabolite mixture from which demethyldictamnine and dictamnic acid were identified by a GC-MS technique. The identity of the two compounds was confirmed by synthesis. Other metabolites were characterized by their mass spectra only. The pattern of the metabolites suggested a possible pathway of metabolism in vitro. We assume that the metabolism of dictamnine is analogous to the metabolism of the related 8-methoxypsoralen and takes place via an unstable epoxide and subsequent oxidative opening of the furan ring. Direct evidence for the formation of an epoxide was, however, not obtained. The identified compounds as well as some putative metabolites were shown to be non-mutagenic in Salmonella typhimurium TA98 except for 8-hydroxydictamnine, which, however, was not detected in the metabolite mixture.

Full text of DOI:10.1093/mutage/14.2.181

Mutagenesis vol.14 no.2 pp.181–185, 1999
Microsomal metabolism of dictamnine: identification of
metabolites and evaluation of their mutagenicity in Salmonella
typhimurium
Bernhard Klier and Oskar Schimmer1
demethyldictamnine and dictamnic acid. The structures of
other metabolites were proposed from their mass spectra. The
identified metabolites and some related compounds were tested
for mutagenicity in S.typhimurium TA 98.
Institut fur Botanik und Pharmazeutische Biologie der Universita¨t Erlangen-
¨
Nurnberg, Staudtstraβe 5, D 91058 Erlangen, Germany
¨
Incubation of the furoquinoline alkaloid dictamnine with
microsomal fractions from phenobarbital-induced rat liver
resulted in a metabolite mixture from which demethyl-
dictamnine and dictamnic acid were identified by a GC-
MS technique. The identity of the two compounds was
confirmed by synthesis. Other metabolites were character-
ized by their mass spectra only. The pattern of the metabol-
ites suggested a possible pathway of metabolism in vitro.
We assume that the metabolism of dictamnine is analogous
to the metabolism of the related 8-methoxypsoralen and
takes place via an unstable epoxide and subsequent oxidat-
ive opening of the furan ring. Direct evidence for the
formation of an epoxide was, however, not obtained. The
identified compounds as well as some putative metabolites
were shown to be non-mutagenic in Salmonella typhimurium
TA98 except for 8-hydroxydictamnine, which, however, was
not detected in the metabolite mixture.
Materials and methods
General
1H-NMR. 1H-NMR were run at 360 and 90 MHz, respectively, in DMSO-d6,
CDCl3 and CD3OD. The internal standard was TMS.
TLC was performed on pre-coated HPTLC plates (silica gel 60 F254
using the solvent system toluene/EtOAc/formic acid (5:4:1). Detection: UV,
anisaldehyde, Dragendorffs reagent.
)
Gas chromatography. For the reference compounds a Vega 2 model with a
DB5 column (25 µm film thickness) was used. The temperature programme
was 80–280°C, 6°C/min. For metabolites a DB1-30W column was employed,
with a temperature program of 150–300°C, 6°C/min.
Gas chromatographic profiles. The retention indices (RI) of the individual
metabolites were determined with a series of alkanes (C18–C24) as standards.
The RI of dictamnine was 1881.5. Only those peaks were analysed which
were formed after incubation of dictamnine with microsomes but not in
control incubations (dictamnine without microsomes or dictamnine plus
inactivated microsomes).
Mass spectrometry. The EI ionization spectra were recorded with a Finnigan
model 4500 and MAT 8930. The ionization energy was 70 eV and the source
temperature was 120°C.
N-Methyl-N-trimethylsilyl trifluoroacetamide (MSTFA reagent) was pur-
chased from Macherey-Nagel (Duren, Germany).
¨
Introduction
The furoquinoline alkaloid dictamnine is very common within
the family Rutaceae. It is the main alkaloid in the roots of
Dictamnus albus and responsible for the mutagenicity of
extracts derived from the crude drug (Mizuta and Kanamori,
1985). Dictamnine was also reported to be a phototoxic and
photomutagenic compound (Pfyffer et al., 1982; Schimmer and
Reference compounds
¨
Dictamnine. Dictamnine was isolated from Dictamni radix (Muggenburg,
Hamburg, Germany). Dried powdered plant material was extracted with
petroleum ether (boiling point 30–60°C) for 48 h in a Soxhlet apparatus.
The extracts were collected and concentrated under reduced pressure. The
concentrate was extracted exhaustively with 3 N HCl. The aqueous acidic
solution was alkalized with NaOH, adjusted to pH 10 and then extracted with
CH2Cl2. The organic phases were collected and evaporated under reduced
pressure. The residue was suspended in a small amount of acetone, centrifuged
and the clear supernatant was chromatographed on a Sephadex LH20 column
with acetone as eluent. The acetone fractions were chromatographed on silica
gel plates. Fractions which showed blue fluorescent spots were collected and
dried. For further purification the residue was again chromatographed on a
Sephadex LH20 with MeOH/water (99:1) as the eluent. This resulted in the
isolation of dictamnine (43 mg/kg crude drug) and of robustine (0.96 mg/kg).
The identity and purity of dictamnine was confirmed by GC-MS analysis.
The data were in agreement with the literature (Paulini et al., 1989).
Kuhne, 1991). It participates in the severe skin phototoxicity of
¨
the plant (Schempp et al., 1996).
Dictamnine mutagenicity has been studied in detail using
Salmonella typhimurium TA98, TA100, TA1535 and TA1538.
The results were reported in several papers (Paulini et al.,
1987, 1989; Hafele and Schimmer, 1988). A relatively strong
activity was detected in experiments with S9 mix from rat
liver. The highest number of revertants was induced by
¨
¨
microsomes from phenobarbital-induced liver (Hafele and
Schimmer, 1988). From inhibitor experiments the authors
concluded that dictamnine was activated by cytochrome P450.
In view of other results, dictamnine N-oxide was discussed as
a possible ultimate mutagen.
On the other hand, the photobiological activity of dictamnine
was shown to be connected with the reactive furan double bond,
which reacts covalently with DNA under UV-A irradiation
conditions (Pfyffer et al., 1982). It seemed worthwhile to test
the possibility that the furan ring could also be involved in
metabolic activation, forming an epoxide structure as the active
intermediate instead of the N-oxide.
8-Hydroxydictamnine (robustine). The extraction and isolation procedures
have been described in the dictamnine section. The UV, MS and 1H-NMR
data and the melting point are in agreement with the literature (Chang et al.,
1976). MS data of the MSTFA derivative: m/z 287 (Mϩ), 272 (100), 257,
242, 227, 214, 201, 186, 156, 136, 100, 73, 45.
7-Hydroxydictamnine (confusameline). The compound was a generous gift
from Prof. F.Tillequin (Universite´ Rene´ Descartes, Paris, France). UV and
MS data and melting point agree with the literature (Johns et al., 1968). 1H-
NMR δ (p.p.m.), 8.19 (1 H, d, H-5, J ϭ 9 Hz), 7.57 (1 H, d, H-2, J ϭ 3 Hz),
7.07 (1 H, dd, H-6, J1 ϭ 9 Hz, J2 ϭ 2.5 Hz), 7.05 (1 H, d, H-3, J ϭ 3 Hz),
7.31 (1 H, d, H-8, J ϭ 2.5 Hz), 5.35 (1 H, s, OH), 4.44 (3H, s, OMe-4). MS
data of the MSTFA derivative, m/z 287 (Mϩ, 100), 272, 257, 242, 228, 200,
186, 149, 136, 102, 73, 45.
Metabolites resulting from incubation with rat liver micro-
somes were analysed by a GC-MS technique and tested
for mutagenicity. This paper describes the identification of
1To whom correspondence should be addressed. Tel: ϩ49 9131 852 8251; Fax: ϩ49 9131 852 8243
© UK Environmental Mutagen Society/Oxford University Press 1999
181

B.Klier and O.Schimmer
Demethyldictamnine. Fifty milligrams of dictamnine were dissolved in 3 ml
HBr ϩ HOAc (1 ϩ 1) and heated under reflux while being stirred for 1 h
(bath temperature 110°C). The reaction mixture was diluted with 20 ml of
distilled water. Then the solution was made alkaline with aqueous sodium
hydroxide (6 N) and extracted with diethylether to remove traces of unreacted
dictamnine. After acidification with 3 N HCl the reaction product was extracted
with diethylether. On removal of the solvent in vacuo a white solid was
obtained. The solid was purified by column chromatography (Sephadex LH20,
MeOH) yielding 18.5 mg (36%) of demethyldictamnine, melting point 244°C.
Data evaluation. The protein content was measured by the method of Lowry
et al. (1951). The cytochrome P450 content was estimated by the method of
Omura and Sato (1964). The values presented above were the mean of eight
individual determinations of a pooled fraction from four rat livers. The data
presented in Tables I and II are the mean of eight plates Ϯ SD from
two separate experiments each performed with quadruplicate plates. The
interexperimental variation was controlled using the F-test (P ϭ 0.05), which
established a homogeneity of variance.
UV (MeOH) λmax ϭ 238, 256, 330, 340 nm; UV (MeOH ϩ HCl) λmax
ϭ
Toxicity. Toxicity was evaluated by microscopic control of the background
238, 310, 326, 338 nm; UV (MeOH ϩ NaOH) λmax ϭ 208, 258, 336,
350 nm. 1H-NMR (90 MHz, in DMSO) δ (p.p.m.), 10.61 (1 H, br, s, NH),
8.26 (1 H, m, H-5), 7.39 – 7.65 (3 H, m, H-2, H-6 or H-7, H-8), 7.15 (1 H,
ddd, H-7 or H-6, J1 ϭ 5 Hz, J2 ϭ 4 Hz, J3 ϭ 1.4 Hz), 6.97 (1 H, d, H-3,
J ϭ 1.8 Hz). MS, m/z (%) 186 (9), 185 (100, Mϩ) 156 (10), 129 (31), 128
(32), 102 (15), 101 (9), 92 (7), 77 (7), 76 (11), 75 (7). MS (MSTFA derivative)
m/z, 257 (100, Mϩ), 242, 226, 214, 184, 140, 113, 73, 45.
lawn.
Microsomal incubations and preparation of the metabolite fraction. PB-
induced microsomes showed stronger activation capacities for dictamnine in
TA98 than uninduced or MC-, BNF- and CLO-induced fractions (Table I).
Therefore, the in vitro experiments were performed with microsomal fractions
obtained from PB-induced rat liver, which were adjusted to 10 mg protein/ml
and 2 nmol cytochrome P450/mg protein. Dictamnine (5 mg, dissolved in
DMSO) was incubated with 100 ml microsomal mix (10 ml microsomal
fraction) for 1 h at 37°C. Incubations were terminated by the addition of 100
ml of ice-cold acetone. The acetone was removed and the residue was frozen
in liquid nitrogen. The frozen mixture was stored dry at –20°C and dissolved
in water. The water extract was chromatographed on a RP18 column using
successively water, water/methanol (4:6) and methanol as eluents. The three
fractions were evaporated to dryness and MSTFA added for derivatization.
The derivatized metabolites were then analysed by a GC-MS technique.
Dictamnic acid. Dictamnic acid was synthesized after the method described
by Monkovic et al. (1967). Yield 20.8% (5.2 mg), melting point 255°C. UV
(MeOH) λmax ϭ 229, 275, 277, 314 nm. 1H-NMR (360 MHz, CD3OD) δ
(p.p.m.), 8.03 (1 H, br, d, H-5, J ϭ 8.5 Hz), 7.65 (1 H, dd, H-6 or H-7, J1 ϭ
J2 ϭ 7.5 Hz, J3 ϭ 1.5 Hz), 7.36 (1 H, br, d, H-8, J ϭ 8.5 Hz), 7.33 (1 H,
ddd, H-7 or H-6, J1 ϭ J2 ϭ 8 Hz, J3 ϭ 1 Hz), 4.22 (1H, s, OMe-4). MS of
the MSTFA derivative, m/z 363 (Mϩ), 348, 332, 318, 304, 274, 256, 246
(100), 200, 159, 147, 133, 73, 45.
Dictamnine N-oxide. The N-oxide was synthesized by analogy to the method
published by Craig and Purushothaman (1970). Yield 15% (7.5 mg), melting
point 90°C (decomposition). UV (MeOH) λmax ϭ 222, 252, 355 nm. UV
(MeOH ϩ HCl) λmax ϭ 210, 243, 335 nm. 1H-NMR (360 MHz, CDCl3) δ
(p.p.m.), 8.63 (1 H, d, H-5, J ϭ 8.5 Hz), 8.41 (1 H, br, d, H-8, J ϭ 8.5 Hz),
8.04 (1 H, d, H-2, J ϭ 3 Hz), 7.97 (1 H, ddd, H-6 or H-7, J1 ϭ J2 ϭ 8 Hz,
J3 ϭ 1.5 Hz), 7.69 (1 H, ddd, H-7 or H-6, J1 ϭ J2 ϭ 7.5 Hz, J3 ϭ 1.5 Hz),
7.61 (1 H, d, H-3, J ϭ 3 Hz), 4.56 (3H, s, OMe-4). MS, m/z (%) 215 (Mϩ,
3), 199 (100), 185 (13), 184 (60), 156 (32), 128 (22), 101 (13), 76 (14).
Results
¨
Hafele and Schimmer (1988) reported that dictamnine can be
activated best with microsomes from PB-induced rat liver. The
results presented here confirm this tendency by comparing the
effect based on the number of revertants. A statistically
significant effect of PB-derived microsomes was observed only
in experiments with 0.1 nmol cytochrome P450/plate. With
higher cytochrome doses no significant differences between
PB-, MC- and BNF-derived microsomes were obtained (Table
I). By taking all the results into consideration, PB-induced
microsomes were used to study the microsomal metabolism
of dictamnine. Liver microsomes from PB-pretreated rats
were incubated with dictamnine. Figure 1 shows a typical
chromatogram of extracts obtained after incubation with
dictamnine.
Mutagenicity experiments
Chemicals. Phenobarbital (PB) was obtained from Bayer (Leverkusen, Ger-
many), 3-methylcholanthrene (MC), β-naphthoflavone (BNF), clofibrate
(CLO), corn oil and 2-aminoanthracene (2-AA) were obtained from Sigma-
Chemie (Deisenhofen, Germany). NADP and isocitrate dehydrogenase came
from Boehringer (Mannheim, Germany). The other chemicals were purchased
from E.Merck (Darmstadt, Germany).
Assay. The assays were performed by the plate incorporation method (Maron
and Ames, 1983) with S.typhimurium TA98.
Microsomal metabolism resulted in the formation of various
peaks eluting with the same retention times as the reference
compounds. The peaks were analysed by mass spectra. The
spectra were compared with the fragmentation pattern of the
reference compound.
S9 fraction. Liver S9 fractions were prepared from male Wistar rats (200–
300 g body wt) which had been treated i.p. daily with PB (70 mg/kg), MC
(25 mg/kg), BNF (80 mg/kg) or CLO (400 mg/kg) for 3 days. The control
group received only corn oil (0.5 ml) and was regarded as uninduced (UN).
Four days after the first injection the animals were killed by cervical
dislocation. The animals were fasted for 24 h prior to killing. S9 fractions
were prepared by a 20 min centrifugation at 9000 g. The protein contents of
the S9 fractions were 30 (PB), 30 (MC), 32.2 (UN), 30.8 (BNF) and 37.9
(CLO) mg/ml, respectively, and was adjusted to 30 mg/ml. The cytochrome
P450 levels were 0.37 Ϯ 0.01 (PB), 0.15 Ϯ 0.005 (MC), 0.11 Ϯ 0.005 (UN),
0.14 Ϯ 0.005 (BNF) and 0.19 Ϯ 0.005 (CLO) nmol/mg protein, respectively.
Figure 2 shows the EI spectrum of the demethyldictamnine
peak derived from dictamnine along with that of authentic
demethyldictamnine. The two spectra are strikingly similar,
with m/z 257 as the molecular ion peak of the MSTFA
derivative and intense fragments at m/z 242 and 214; m/z 73
was identified as a fragment of the MSTFA reagent. A second
metabolite was identified as dictamnic acid. The mass spectrum
of the derivatized substance showed a molecular ion peak at
m/z 363 and fragments at m/z 348 [M-CH3], 332 [M-OCH3],
318 [M-3CH3], 246 [M-COOTMS] and 274 [M-OTMS]. The
authentic MSTFA derivative showed a very similar pattern of
fragments. A further metabolite, M287, was identified as a
hydroxydictamnine, but the position of the hydroxyl group
could not be deduced from the data. In the mass spectrum a
prominent Mϩ peak at m/z 287 and a major peak at m/z 272
[M-CH3] were observed. Two other hydroxy derivatives were
analysed by MS for comparison. 8-Hydroxydictamnine
(robustine) showed similar peaks at m/z 272 and 257 but a
minor molecular ion peak at m/z 287. 7-Hydroxydictamnine
(confusameline) showed major peaks at m/z 287 and 272 but
S9 mix. One millilitre of the standard activation mixture contained 0.1 ml of
S9 fraction, 0.1 ml of 70 mM MgSO4, 0.1 ml of 40 mM NADP, 0.1 ml of
200 mM D,L-isocitrate and 0.6 ml of 100 mM phosphate buffer (pH 7.4).
Microsomal fraction. A part of the S9 fraction was centrifuged for 1 h at
100 000 g. The supernatant was used as a cytosolic fraction (26 mg protein/
ml). The pellet was resuspended in phosphate buffer and again centrifuged.
The resulting pellet was resuspended in phosphate buffer and analysed for its
protein and cytochrome P450 content. The data on the cytochrome P450
content (nmol/mg protein) were as follows: 0.84 Ϯ 0.04 (UN); 2.09 Ϯ 0.13
(PB); 1.41 Ϯ 0.02 (MC); 1.54 Ϯ 0.09 (CLO); 1.24 Ϯ 0.03 (BNF).
Microsomal mix. Microsomal fraction (0.1 ml) and isocitrate dehydrogenase
(0.05 ml) were added to the standard activation mixture (see S9 mix) for
microsomal activation.
182

Microsomal metabolism of dictamnine
Table I. Metabolic activation capacity of the various microsomal fractions (indicator Salmonella typhimurium TA98)
Substrate
Cyt-P 450 content
(nmol/plate)
Revertants/plate (ϮSD)a
UN microsomes
PB microsomes
MC microsomes
BNF microsomes
CLO microsomes
Dictamnine (250 nmol)
0.1
200 Ϯ 15
624 Ϯ 43
960 Ϯ 89
4568 Ϯ 210
314 Ϯ 17b
950 Ϯ 51
1586 Ϯ 89
1761 Ϯ 97
238 Ϯ 17
844 Ϯ 68
1381 Ϯ 42
1665 Ϯ 61
247 Ϯ 22
868 Ϯ 41
1432 Ϯ 109
1339 Ϯ 46
97 Ϯ 7
246 Ϯ 15
414 Ϯ 32
1658 Ϯ 125
0.25
0.50
0.50
2-AA (1 µg/plate)
aSpontaneous number of revertants were from 17 to 37/plate.
bSignificantly different (t-test): BNF microsomes, P Ͻ 0.05; MC microsomes, P Ͻ 0.01; UN and CLO microsomes,P Ͻ 0.001.
Table II. Mutagenicity of dictamnine and its putative metabolites in vitro in Salmonella typhimurium TA98
Dose
Revertants/plate (ϮSD)a
(nmol/plate)
Dictamnine
Demetyldictamnine
–S9 ϩS9
Dictamnic acid
8-Hydroxydictamnine
7-Hydroxydictamnine Dictamnine N-oxide
–S9
ϩS9
26 Ϯ 5
–S9
ϩS9
–S9
ϩS9
26 Ϯ 3
–S9
ϩS9
–S9
ϩS9
0b
5
25
50
100
250
500
24 Ϯ 4
21 Ϯ 2
31 Ϯ 4
27 Ϯ 4
28 Ϯ 5
26 Ϯ 5
51 Ϯ 3
22 Ϯ 3 38 Ϯ 6
23 Ϯ 5 32 Ϯ 6
25 Ϯ 7 32 Ϯ 5
24 Ϯ 7 32 Ϯ 6
23 Ϯ 6 38 Ϯ 6
20 Ϯ 1 38 Ϯ 9
22 Ϯ 4
31 Ϯ 5
27 Ϯ 6
31 Ϯ 4
28 Ϯ 4
32 Ϯ 3
28 Ϯ 3
37 Ϯ 5
37 Ϯ 2
34 Ϯ 5
37 Ϯ 5
42 Ϯ 7
38 Ϯ 5
40 Ϯ 3
21 Ϯ 6
22 Ϯ 4
20 Ϯ 4
19 Ϯ 4
26 Ϯ 3
30 Ϯ 4
24 Ϯ 4 26 Ϯ 5
85 Ϯ 14
298 Ϯ 28
548 Ϯ 45
1051 Ϯ 83
1557 Ϯ 67
188 Ϯ 15
17 Ϯ 2
16 Ϯ 3
19 Ϯ 4
20 Ϯ 5
18 Ϯ 6
28 Ϯ 3
32 Ϯ 8
32 Ϯ 7
37 Ϯ 5
46 Ϯ 4
744 Ϯ 82
26 Ϯ 2
34 Ϯ 2
1101 Ϯ 75
27 Ϯ 6
24 Ϯ 4
20 Ϯ 2
24 Ϯ 1
37 Ϯ 5
57 Ϯ 12
tc
t
1863 Ϯ 93c 20 Ϯ 5
46 Ϯ 6
40 Ϯ 9
t
2 Ϯ 2c 10 Ϯ 10c 20 Ϯ 7 56 Ϯ 4c
aData are the mean of eight plates from two separate experiments each performed with quadruplicate plates.
bSolvent control, DMSO (50 µl).
cToxic.
Fig. 1. Typical gas chromatogram of extracts of dictamnine incubated with
microsomes. Metabolite M257 at relative retention time 406 min is
indicated. This corresponds to a RI of 1945.2 (see Material and Methods).
its retention time was significantly higher than that of the
metabolite (RI ϭ 2266.3 versus 2213.0). The other metabolites
(M451, 465, 349, 363b and 377) were characterized only
by their mass spectra, since no reference compounds were
available.
Metabolite M451 was a major product of the in vitro
incubation of dictamnine. It showed only a weak signal of the
molecular ion at m/z 451. The fragment m/z 348 showed an
intense peak and may be explained by the loss of CH2OSi
(CH3)3; a minor peak at m/z 260 was possibly due to the loss
of a further 88 mass units [Si(CH3)4], followed by formation
of a heterocyclic ring. Fragments m/z 147 and 73 were
derived from the MSTFA reagent. Metabolite M465 showed
Fig. 2. EI ionization mass spectra of (A) demethyldictamnine, the MSTFA
derivative from an extract of dictamnine incubated with microsomes, and
(B) the MSTFA derivative of authentic demethyldictamnine.
183

B.Klier and O.Schimmer
a fragmentation pattern similar to M451 but with no detectable
molecular ion peak. Characteristic peaks were shown at m/z
348 (M-COOTMS) and 260 (analogous to M451). M349
showed a fragmentation pattern with m/z 349 (Mϩ), 334 (M-
CH3), 318 (M-OCH3), 304 (M-3CH3), 260 (M-OTMS), 246
(M-CH2OTMS), 244, 230, 147, 133 and 73. M363b showed
a molecular ion at m/z 363, but was characterized by a different
fragmentation pattern and different retention time compared
with dictamnic acid. The major peaks were at m/z 348 (M-
CH3), 318 (M-3CH3), 260 (M-OTMS), 246, 244, 232, 230,
147, 103, 73 and 45. Metabolite M377 was identified only
from the urine of rats treated orally with dictamnine. The MS
showed a weak molecular ion at m/z 377 and major peaks at
m/z 362 (M-CH3), 348, 334, 318, 260 (M-COOTMS), 246
(M-CH2COOTMS), 244, 230, 200, 156, 147, 73 and 45.
Although this metabolite was not detected in microsomal
metabolism, it fitted well in the degradation scheme proposed
for dictamnine.
Some identified and putative metabolites were synthesized
and evaluated for their mutagenicity in S.typhimurium TA98.
The N-oxide was included because this compound has been
discussed as the eventual ultimate mutagen enzymatically
formed by microsome mix. None of these compounds were
capable of inducing revertants, apart from 8-hydroxydictam-
nine, whose presence in the metabolite mixture could not be
verified (Table II). Thus, the N-oxide and the other compounds
had to be excluded as possible ultimate mutagens.
Discussion
From our results with the various microsomal preparations it
is evident that PB ist the most active inducer of the enzymes
which metabolize dictamnine into a mutagen. This also
becomes clear when we compare the activity on the basis of
the cytochrome P450 content. Since it is known that PB
preferentially induces cytochromes P450 of the 2B subfamily
(McManus and McKinnon, 1991), one of these enzyme forms
may be particularly involved in the bioactivation of dictamnine.
Among the metabolites separated by GC and identified by
MS, neither dictamnine N-oxide nor dictamnine 2,3-oxide were
detected. To test the possibility that the N-oxide acted as the
ultimate mutagen but was not detected because of its instability,
we synthesized the compound and examined it for its muta-
genicity. The negative result clearly demonstrated the inability
of the N-oxide to function as the mutagenic intermediate.
On the other hand, formation of an epoxide as an intermediate
seemed to be a possible alternative, since the furyl double
bond represents a crucial site for chemical reactions. This was
also indicated by the inactivity found in the corresponding
2,3-dihydro derivative (Klier, 1993; data available on request).
The postulated epoxide, however, could not be isolated nor
synthesized, possibly due to its short-lived nature. However,
the formation of furan epoxides is a common and biologically
relevant phenomenon, which often leads to mutagenicity
(Essigmann et al., 1977; Baertschi et al., 1988; Adam et al.,
1990, 1991). The most convincing example came from meta-
bolism experiments with 8-methoxypsoralen. In vivo this
furocoumarin undergoes O-demethylation, hydroxylation and
oxidative cleavage of the furan ring. The metabolism has been
postulated to take place via the 2,3-epoxide (Kolis et al.,
1979; Mays et al., 1987, 1989). Experiments with rat liver
microsomes were, however, not successful. This was possibly
Fig. 3. Proposed pathway of metabolism of dictamnine in vitro. M377 was
detected in urine from rats treated orally with dictamnine. M365 was
detected after microsomal incubation, but not in urine samples.
caused by covalent binding of metabolites to microsomal
proteins (Mays et al., 1987).
The similarity of dictamnine and 8-methoxypsoralen in
chemical structure and reactivity and the pattern of metabolites
led us to postulate an analogous pathway of metabolism
from our results (Figure 3). The demethyl and the hydroxy
derivatives were formed in this, but the quinoline derivatives
dominated as a consequence of oxidative opening of the furan
ring. A few metabolites in the pathway were only identified
tentatively by their mass spectra and have not yet been
confirmed by synthesis. Nevertheless, there is a certain plausib-
ility to the pathway proposed analogous to 8-methoxypsoralen.
Although this pattern is consistent with the idea that the
metabolism of dictamnine is via a furan epoxide, no direct
evidence was found that such an intermediate has been formed
and acted as the ultimate mutagen. Therefore, we cannot
184

Microsomal metabolism of dictamnine
zu einem Mutagen im Salmonella Mikrosomen Test: Struktur-Wirkungs-
Analysen und Isolierung von Metaboliten. Doctoral Thesis University of
exclude the possibility that an unidentified compound, e.g. the
hydroxy derivative, is involved in dictamnine mutagenicity.
M377 was not detected in vitro, but this may be explained
by the finding that microsomal metabolites were formed to a
lesser extent than the metabolites isolated from in vivo experi-
ments (Klier, 1993). Assuming that the furan ring serves as
the reactive site, we must ask why some metabolites with an
intact furan ring cannot be activated by microsomes or S9
mix. In the case of the demethyl derivative the lack of
metabolic activation may be explained by the finding that this
compound can exist in two tautomeric forms, namely 4-
hydroxyfuroquinoline and furoquinolone-4. This is indicated
Erlangen-Nurnberg, Germany.
¨
Kolis,S.J., Williams,T.H., Postma,E.J., Sasso,G.J., Confalone,P.N. and
Schwartz,M.A. (1979) The metabolism of 14C-methoxsalen by the dog.
Drug Metab. Dispos., 7, 220–225.
Lowry,O.H., Rosebrough,N.J., Farr,A.L. and Randall,R.J. (1951) Protein
measurement with the Folin phenol reagent. J. Biol. Chem., 193, 265–275.
Maron,D.M. and Ames,B.N. (1983) Revised methods for the Salmonella
mutagenicity test. Mutat. Res., 113, 173–215.
Mays,D.C., Hecht,S.G., Unger,S.E., Pacula,C.M, Climie,J.M., Sharp,D.E. and
Gerber,N. (1987) Disposition of 8-methoxypsoralen in the rat. Induction of
metabolism in vivo and in vitro and identification of urinary metabolites by
thermospray mass spectrometry. Drug Metab. Dispos., 15, 318–329.
Mays,D.C., Hilliard,J.B., Wong,D.D. and Gerber,N. (1989) Activation of 8-
methoxypsoralen by cytochrome P 450. Enzyme kinetics of covalent binding
and the influence of inhibitors and inducers of drug metabolism. Biochem.
Pharmacol., 38, 1647–1655.
1
by a signal in the H-NMR which splits at room temperature
(Klier, 1993). The quinolone structure may have been favoured
under our treatment conditions. This agrees with the result
that isodictamnine, another quinolone derivative, was shown
to be inactive in TA98 Ϯ S9 mix (Klier, 1993).
McManus,M.E. and McKinnon,R.A. (1991) Measurements of cytochrome
P450 activation of xenobiotics using the Ames Salmonella test. Methods
Enzymol., 206, 501–509.
It is more difficult to explain the different activities of
the two hydroxy derivatives. Whereas 8-hydroxydictamnine
(robustine) showed strong mutagenicity with S9 mix, the 7-
hydroxy derivative and other furoquinolines substituted with
an hydroxy group at C7 were inactive in TA98 (Klier, 1993;
this work). As the position of the hydroxy group in M287 is
not known at present, far-reaching conclusions on its activity
cannot be drawn. However, it seems very unlikely that 8-
hydroxydictamnine is the ultimate mutagen because of its
stability and its absence in the metabolite mixture.
Mizuta,M. and Kanamori,H. (1985) Mutagenic activities of dictamnine and
gamma-fagarine from Dictamni radicis cortex (Rutaceae). Mutat. Res., 144,
221–225.
Monkovic,I., Spencer,I.D. and Plunkett,A.O. (1967) Biosynthesis of
dictamnine. The origin of the quinoline nucleus. Can. J. Chem., 45,
1935–1948.
Omura,T. and Sato,R. (1964) The carbon monoxide-binding pigment of liver
microsomes. J. Biol. Chem., 239, 2370–2385.
Paulini,H., Eilert,U. and Schimmer,O. (1987) Mutagenic compounds in an
extract from Rutae Herba (Ruta graveolens L.). I. Mutagenicity is partially
caused by furoquinoline alkaloids. Mutagenesis, 2, 271–273.
Paulini,H., Waibel,R. and Schimmer,O. (1989) Mutagenicity and structure–
mutagenicity relationships of furoquinolines, naturally occurring alkaloids
of the Rutaceae. Mutat. Res., 227, 179–186.
Pfyffer,G.E., Panfil,I. and Towers,G.H.N. (1982) Monofunctional covalent
photobinding of dictamnine, a furoquinoline alkaloid, to DNA as a target
in vitro. Photochem. Photobiol., 35, 63–68.
Acknowledgements
The authors acknowledge with gratitude the valuable help in the GC-MS
analysis of the metabolite fractions by Dr U.Eilert (Institute of Pharmaceutical
Biology, University of Braunschweig). We are also indebted to Dr R.Waibel
Schempp,C.M., Sonntag,M., Schopf,E. and Simon,J.C. (1996) Dermatitis
¨
¨
(Institute of Pharmacy, University of Erlangen-Nurnberg) for taking and
bullosa striata pratensis durch Dictamnus albus L. (Brennender Busch).
Hautarzt, 47, 708–710.
interpreting the NMR and mass spectra of the reference compounds. We thank
¨
Prof. Estler and Prof. Bocker (Institute of Pharmacology and Toxicology,
Schimmer,O. and Kuhne,I. (1991) Furoquinoline alkaloids as photosensitizers
¨
¨
University of Erlangen-Nurnberg) for their help during the course of the
in Chlamydomonas reinhardtii. Mutat. Res., 249, 105–110.
experiments with inducers. We would also like to thank Prof. Tillequin
(Laboratoire de Pharmacognosie, Faculte de Pharmacie, Universite Rene
Descartes, Paris) for generously providing a sample of 7-hydroxydictamnine
and Prof. H.R.Glatt for critical reading of the manuscript.
Received on April 15, 1998; accepted on November 6, 1998
References
Adam,W., Hauer,H., Mosandl,T., Saha-Mo¨ller,C.R., Wagner,W. and Wild,D.
(1990) Furocoumarin-, naphthofuran- and furoquinoline-annulated 1,2-
dioxetanes: synthesis, thermolysis and mutagenicity. Liebigs Ann. Chem.,
1990, 1227–1236.
Adam,W., Hadjiarapoglou,L., Mosandl,T., Saha-Mo¨ller,C.R. and Wild,D.
(1991) Chemical model studies on the mutagenesis of benzofuran dioxetanes
in the Ames test. Evidence for the benzofuran epoxide as ultimate mutagen.
J. Am. Chem. Soc., 113, 8005–8011.
Baertschi,S.W., Raney,K.D., Stone,M.P. and Harris,T.M. (1988) Preparation
of the 8,9-epoxide of the mycotoxin aflatoxin B1: the ultimate carcinogenic
species. J. Am. Chem. Soc., 110, 7929–7931.
Chang,P.T.O., Cordell,G.A., Aynilian,G.H., Fong,H.H.S. and Farnsworth,N.R.
(1976) Alkaloids and coumarins of Thamnosma montana. Lloydia, 39,
134–140.
Craig,J.C. and Purushothaman,K.K. (1970) An improved preparation of tertiary
amine N-oxides. J. Org. Chem., 35, 1721–1722.
Essigmann,J.M., Croy,R.G., Nadzan,A.M., Busby,W.F., Reinhold,V.N.,
Buchi,G. and Wogan,G.N. (1977) Structural identification of the major
¨
DNA adduct formed by aflatoxin B1 in vitro. Proc. Natl Acad. Sci. USA,
74, 1870–1874.
¨
Hafele,F. and Schimmer,O. (1988) Mutagenicity of furoquinoline alkaloids in
the Salmonella/microsome assay. Mutagenicity of dictamnine is modified
by various enzyme inducers and inhibitors. Mutagenesis, 3, 349–353.
Johns,S.R., Lamberton,J.A. and Sioumis,A.A. (1968) Furoquinoline alkaloids
from Evodia elleryana F. Muell. The structure of evellerine. Aust. J. Chem.,
21, 1897–1901.
Klier,B. (1993) Untersuchung der metabolischen Aktivierung von Dictamnin
185