Human intestinal bacteria mediate reduction of the N-oxides of isoline and monocrotaline to the corresponding parent alkaloids

Article abstract of DOI:10.14233/ajchem.2013.13283

The role of human intestinal bacteria in the biotransformation of pyrrolizidine alkaloid N-oxides was investigated. Two naturally-occurring hepatotoxic pyrrolizidine alkaloids, isoline and monocrotaline and their N-oxides were incubated with a human fecal suspension in an in vitro model, respectively. The metabolites were isolated and purified by chromatographic techniques and identified by the spectral analyses and the metabolic profiles were further analyzed by a specific TLC method. Human intestinal bacteria showed high reduction effects on the pyrrolizidine alkaloid N-oxides but not on the resultant tertiary alkaloids, suggesting that IB may play a partial role in the cyclic conversion between each pyrrolizidine alkaloid and its N-oxide in vivo. This evidence implied the potential risk of the pyrrolizidine alkaloid-containing Chinese medicinal herbs to human health when used in decoctions.

Full text of DOI:10.14233/ajchem.2013.13283

Asian Journal of Chemistry; Vol. 25, No. 4 (2013), 2027-2030
http://dx.doi.org/10.14233/ajchem.2013.13283
Human Intestinal Bacteria Mediate Reduction of the N-Oxides of
Isoline and Monocrotaline to the Corresponding Parent Alkaloids
2,3
4
5
JUN TANG1,2,*, ZHENGTAO WANG , TERUAKI AKAO and MASAO HATTORI
1Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education and Wuhan University
School of Pharmaceutical Sciences, Wuhan 430071, P.R. China
2Department of Pharmacognosy, China Pharmaceutical University, Nanjing 210038, P.R. China
3Institute of Chinese Materia Medica, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, P.R. China
4Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, 2630 Sugitani, Toyama 930-019, Japan
5Institute of Natural Medicine, University of Toyama, 2630 Sugitani, Toyama 930-019, Japan
*Corresponding author: Fax: +86 27 68759850; Tel: +86 27 68759923; E-mail: tangj0205@gmail.com
(Received: 19 December 2011;
Accepted: 12 October 2012)
AJC-12271
The role of human intestinal bacteria in the biotransformation of pyrrolizidine alkaloid N-oxides was investigated. Two naturally-occurring
hepatotoxic pyrrolizidine alkaloids, isoline and monocrotaline and their N-oxides were incubated with a human fecal suspension in an in
vitro model, respectively. The metabolites were isolated and purified by chromatographic techniques and identified by the spectral analyses
and the metabolic profiles were further analyzed by a specific TLC method. Human intestinal bacteria showed high reduction effects on
the pyrrolizidine alkaloid N-oxides but not on the resultant tertiary alkaloids, suggesting that IB may play a partial role in the cyclic
conversion between each pyrrolizidine alkaloid and its N-oxide in vivo. This evidence implied the potential risk of the pyrrolizidine
alkaloid-containing Chinese medicinal herbs to human health when used in decoctions.
Key Words: Isoline N-oxide, Monocrotaline N-oxide, Hepatotoxic pyrrolizidine alkaloids, Human intestinal bacteria, Chinese
medicinal herbs.
some pyrrolizidine alkaloid N-oxides including monocrotaline
N-oxide might be reduced to their toxic parent pyrrolizidine
alkaloids in human liver microsomes1,8 and the N-oxides of
indicine and retrorsine could be reduced by the gut flora in
the rabbit or rats9,10. The previous studies also showed that the
gut may be the major site of the formation of PAs from their
N-oxides3,9. However, it remains unclear about what effects
the human gut flora may have in the conversion between
pyrrolizidine alkaloids and their N-oxides. Accordingly, this
study reported the role of human intestinal bacteria (IB) in the
metabolism of pyrrolizidine alkaloid and its N-oxide, using
two known hepatotoxic pyrrolizidine alkaloids, that is isoline
and monocrotaline, as the examples.
INTRODUCTION
Hepatotoxic pyrrolizidine alkaloids (PAs) are well known
to be a kind of phytotoxins, with a core structure containing
1,2-unsaturated necines, e.g., retronecine or otonecine and
necic acids1. Isoline and monocrotaline are two retronecine-
based cyclic diester pyrrolizidine alkaloids, showing high hepato-
toxicity and/or pneumotoxicity to many animal species2,3.
Meanwhile, they were found to be abundant in some Chinese
medicinal herbs, for example, a high level of isoline was found
in Ligularia duciformis (C. Winkl.) Hand.-Mazz., which substi-
tutes Asteris Radix as an antitussive and expectorant agent4;
monocrotaline could be highly accumulated in seeds of
Crotalaria assamica Benth, which have a special use in treatment
of skin cancer and leukemia5. It is interesting to note that these
retronecine-type pyrrolizidine alkaloids may co-exist with their
corresponding N-oxides in Chinese medicinal herbs, while
more polar N-oxides and less tertiary alkaloids may be ingested
together in decoctions by patients in Chinese medical practice1,6.
No matter what functions these N-oxides may have in the
host plants, they may represent the detoxification products of
their parent pyrrolizidine alkaloids in vivo6,7. On the other hand,
EXPERIMENTAL
Isoline and monocrotaline were isolated from Ligularia
duciformis and Crotalaria assamica, respectively and their
N-oxides were semi-synthesized as described previously4,11,12
.
All chemicals and solvents were at least of AR grade. A Tabai
anaerobic incubator EAN-140 (Tabai Co., Osaka, Japan) was
used for anaerobic incubation. General anaerobic medium
(GAM) was purchased from Nissui Co. (Tokyo, Japan). K+-

2028 Tang et al.
Asian J. Chem.
phosphate buffer (pH 7.26) was prepared as reported previously13.
The modified Ehrlich reagent was prepared by dissolving
4-dimethylaminobenzaldehyde (2.0 g) with boron trifluoride
ether complex (2 mL) in absolute ethanol (100 mL) freshly or
kept in dark place until use. TLC was performed on a precoated
silica gel 60 F254 plate (0.25 mm, Merck) using chloroform-
methanol-28 % ammonia solution as developing system in a
ratio from 8:2:0.1 to 7:3:0.1 (v/v). Column chromatography
was carried out on silica gel BW-820MH or ODS (Fuji Silysia
Co., Nagoya, Japan). Preparative TLC proceeded on silica gel
60 F254 plates (0.5 mm, Merck) and the strips were detected
For the identification of metabolites, a scaled-up incuba-
tion was performed with a human fecal suspension (50 mL)
containing each N-oxide (0.5 mM) at 37 ºC for 5 h. The final
mixtures were then extracted with n-butanol and separated
through repeated column chromatography and/or preparative
TLC.
RESULTS AND DISCUSSION
Characterization of the metabolites of two pyrrolizi-
dine alkaloid N-oxides:After the scaled-up incubations with
two PA N-oxides, metabolite 1 (5 mg) and 2 (3 mg) were
obtained with a rate of 51 and 38 %, respectively. As shown
below, metabolite 1 was obtained from isoline N-oxide as a
white needle. The EIMS showed a molecular ion peak [M]+ at
m/z 395 as a base. Metabolite 2 was obtained from
monocrotaline N-oxide as a fan-shaped crystal. The EIMS
showed a molecular ion peak [M]+ at m/z 325. The MS
fragmentation pattern for each compound was similar to that
of isoline and monocrotaline, respectively. Meanwhile, both
metabolites gave the same characteristic ions of a retronecine-
type necine base at m/z 136, 120, 93 and 80. The 1H and 13C
NMR spectral data of metabolites 1 and 2, were in agreement
with those reported for isoline14 and for monocrotaline15,
respectively. The co-TLC analyses with authentic samples also
confirmed 1 and 2 as isoline and monocrotaline, respectively
(Fig. 1). The lower yield rate of 2 may be ascribed to its higher
hydrophilicity and poor partition in n-butanol.
1
under a UV light or with iodine. H and 13C NMR spectra
were measured with a Varian Unity Plus 500 (1H, 500 MHz;
13C, 125 MHz) spectrometer (Varian, CA) in CDCl3 or CD3OD.
EIMS was measured with a JEOL JMS-GC mate mass spectro-
meter (Jeol Co., Tokyo, Japan).
Preparation of a human fecal suspension:A fresh human
fecal sample (1 g, collected from different part of fecal mass),
obtained from a healthy subject (male, Chinese, 31 years old
or Egyptian, 46 years old) with informed consent, was homo-
genized in GAM broth (5.9 % water solution, 20 mL) or K+-
phosphate buffer (50 mM, 20 mL) under anaerobic condition
and the sediments were removed by filtration to give a human
fecal suspension. The suspensions were used as incubation
matrices in the following experiments.
Preparation of pyrrolizidine alkaloids standard stock
solutions: Isoline and monocrotaline were dissolved in water
with a suitable volume of 0.2 M hydrochloric acid followed
by neutralization with 0.2 M sodium hydroxide and isoline
and monocrotaline N-oxides were dissolved in water directly
to give the standard stock solutions containing 50 mM of each
compound, respectively.
Metabolite 1 A white needle (hexane/acetone), m.p. 168-
169 ºC. EIMS m/z (%): 395 (M+, 100), 366 (M+-CH3CH2, 8),
335 (M+-CH3COOH, 15), 320 (M+-CH3COOH-CH3, 10), 309,
308 (M+-CH3CO-CO2, 98), 291 (309-H2O, 11), 236 (308-
CH3COCH2CH3, 16), 220 (308-CH3COCHCH3OH, 50), 138,
136, 120, 118, 108, 95, 93, 83, 80. 1H NMR (CDCl3, 500 MHz)
δ 6.18 (1H, br s, H-2), 3.97 (1H, dd, H-3α), 3.33 (1H, m, H-
3β), 3.31 (1H, m, H-5α), 2.45 (1H, m, H-5β), 2.29 (1H, dd,
H-6α), 2.12 (1H, m, H-6β), 5.06 (1H, t, H-7), 4.24 (1H, br m,
H-8 ), 5.17 (1H, d, H-9α), 4.18 (1H, d, H-9β), 1.77 (1H, m,
H-13), 1.59 (1H, m, H-14α), 1.39 (1H, dd, H-14β), 1.38 (3H,
s, H-18), 1.21 (3H, d, H-19), 1.59 (2H, m, H-20), 0.80 (3H, t,
H-21), 2.03 (3H, s, CH3CO). 13C NMR (CDCl3, 125 MHz) δ:
131.2 (C-1), 135.9 (C-2), 63.4 (C-3), 53.2 (C-5), 34.5 (C-6),
76.8 (C-7), 77.4 (C-8), 60.6 (C-9), 172.2 (C-11), 83.5 (C-12),
37.2 (C-13), 39.4 (C-14), 78.8 (C-15), 176.1 (C-16), 14.8
(C-18), 15.8 (C-19), 33.3 (C-20), 7.4 (C-21), 170.0 (COCH3),
21.1 (COCH3).
Incubation of pyrrolizidine alkaloids with a human
fecal suspension: All incubations were conducted according
to a known in vitro model previously established in this labo-
ratory13 with some modifications. Briefly, to the human fecal
suspension (5 mL), standard solutions (50 µL) of the PAs
(including N-oxides) were added in aliquots, respectively and
then incubated at 37 ºC in the anaerobic incubator for more
than 6 h. Meanwhile, various controls including a control with-
out PA, a control in GAM broth without feces or that under
aerobic condition were examined in parallel. A 100 µL of
portion was taken out at intervals of at least 1 h and extracted
with 100 µL of n-butanol (saturated with water and adjusted
to pH 8-9 with 28 % ammonia solution). After centrifugation
at 10,000 g for 3 min, 5 µL of the n-butanol layer, the corres-
ponding aqueous layer as well as different standard solution
was applied to TLC, respectively. Afterwards, three steps
proceeded for the visualization and detection of the substrates
and possible metabolites, that is, first oxidized with iodine
and/or 30 % hydrogen peroxide (only for parent alkaloids),
then dehydrogenated with acetic anhydride and finally reacted
with Ehrlich reagent. All steps need heating except iodine
exposure. The appearance of a distinct magenta colour (Ehrlich
colour) was regarded as the positive reaction. In this condi-
tion, the limits of detection could be 0.1 µg for isoline and
0.3 µg for monocrotaline. All experiments were conducted in
triplicate.
Metabolite 2 A white fan-shaped crystal (chloroform),
m.p. 202 ºC. EIMS m/z (%): 325 (M+, 23), 281 (M+-CO2, 7),
237 (281-CH3COH, 75), 236 (100), 193 (237-CH3COH, 14),
165 (193-CH3CH, 6), 137, 136, 120, 119, 94, 93, 80. 1H NMR
(CDCl3, 500 MHz) δ: 6.04 (1H, br.s, H-2), 3.90 (1H, d, H-
3α), 3.48 (1H, dd, H-3β), 3.20 (1H, m, H-5α), 2.59 (1H, dd,
H-5β), 2.08 (2H, m, H-6), 5.05 (1H, m, H-7), 4.38 (1H, br m,
H-8), 4.90 (1H, d, H-9α), 4.69 (1H, d, H-9β), 2.80 (1H, dd,
14-H), 1.44 (3H, s, H-17), 1.35 (3H, s, H-18), 1.22 (3H, d, H-
19). 13C NMR (CDCl3, 125 MHz) δ: 132.7 (C-1), 134.3 (C-2),
61.3 (C-3), 53.6 (C-5), 33.5 (C-6), 75.1 (C-7), 76.8 (C-8),
60.5 (C-9), 174.0 (C-11), 76.7 (C-12), 78.7 (C-13), 44.3 (C-
14), 173.5 (C-15), 21.9 (C-17), 17.6 (C-18), 13.6 (C-19).

Vol. 25, No. 4 (2013)
Human Intestinal Bacteria Mediate Reduction of the N-Oxides of Isoline and Monocrotaline 2029
Fig. 1. Reduction of isoline or monocrotaline N-oxide to the corresponding tertiary alkaloid by intestinal bacteria and other metabolic pathways mediated
by liver enzymes
Metabolic profiles of two pyrrolizidine alkaloid N-
oxides and their parent alkaloids by human intestinal
bacteria: The biotransformation of two hepatotoxic PAs and
their N-oxides were conducted in a previously described model
for in vitro metabolism involved by intestine bacteria, which
could better simulate an in vivo condition13. Moreover, a specific
TLC method employing Ehrlich reagent was developed for
the examination of PAs and their metabolites, whose specificity
was based on the unique pyrrolic derivatives from PAs or their
metabolites and the associated nucleophilic reaction with the
Ehrlich reagent.As a result, isoline N-oxide and monocrotaline
N-oxide were completely transformed within 4.5 and 3.0 h
into their tertiary bases, respectively (Fig. 2A, C). The two
metabolites had higher Rf values (0.77 and 0.62) than their
corresponding N-oxide substrates (0.47 and 0.18), showing a
remarkable loss in the polarity after conversion. TLC results
from the various controls indicated that the change was mainly
mediated by IB. Among them, the control without PA showed
no other spots with the same Rf values and colours as either
N-oxide substrates or their metabolites (Fig. 2E). The control
without IB showed a slight decrease of two N-oxide substrates
and the corresponding formation of parent PAs, however, this
process was so slow that large amount of N-oxides remained
unchanged even after more than 10 h, in which both N-oxides
were almost stable within the first 3 h incubation, whereas at
least 4 h under aerobic condition (Fig. 2B, D).
Conversely, both isoline and monocrotaline were found
unchanged during 120 h incubation with human intestine
bacteria. By TLC, no other spots except PA substrates were
detected with positive Ehrlich colour (Fig. 2E). Moreover, this
study has also examined the different human race and rat strain
and/or used clivorine (an otonecine-type PA) as substrate and
showed same results too (data not shown here). In contrast to
the liver in which monocrotaline, isoline and clivorine can be
bioactivated or detoxified3,7,16, the gut flora may not be a meta-
bolic site for these tertiary alkaloids (Fig. 1).
Taken together, the resultant metabolites of two PA N-
oxides by human IB were the less polar but more stable
tertiary alkaloids, which may more likely be re-absorbed from
the intestinal tract or through enterohepatic circulation
(Scheme-I). The metabolic rates showed somewhat different
for two N-oxides by TLC, suggesting that this conversion may
be substrate-dependent. Even though two substrates showed a
slight chemical decomposition under anaerobic condition, they
did not change so extensively and rapidly as in the presence
of IB prior to excretion. These evidences indicated that human
IB play an essential role in the rapid and irreversible reduction
of PA N-oxides in the intestinal tract, which supported that
the detoxification of PAs via N-oxidation may be reversible
and oral ingestion of PAs and their N-oxides through a decoc-
tion of Chinese herbs may pose a potential risk on human
health1,6,17
.

2030 Tang et al.
Asian J. Chem.
Scheme-I: Proposed cyclic conversion between PA and its N-oxide
through enterohepatic circulation in vivo and a partial role of
the intestinal bacteria. The N-oxides of isoline and
monocrotaline derived from two hepatotoxic Chinese herbs
could be completely and rapidly reduced to their parent
alkaloids by human intestinal bacteria, which may be re-
absorbed into the liver together with the originally contained
tertiary PAs and suffered further bioactivation. Some N-oxides
formed in liver may be excreted into the gut again with biles
ACKNOWLEDGEMENTS
Fig. 2. Representative TLC profiles of metabolism of isoline, monocrotaline
and their N-oxides by human intestinal bacteria. (A) Incubation of
isoline N-oxide with human fecal suspension. 1-6 n-butanol extracts
after incubation for 0, 1, 3, 4.5, 6, 8.5 h; 7 isoline; 8-13 the
corresponding H2O layers after n-butanol extraction; 14 isoline N-
oxide; (B) Control incubation of isoline N-oxide in GAM broth
without feces. 1,7 n-butanol extract and the corresponding H2O layer
of control without PA; 2-6 n-butanol extracts after incubation for
0, 3, 7, 11, 19 h; 8-12 the corresponding H2O layers after n-butanol
extraction; 13 isoline N-oxide; 14 isoline; (C) Incubation of
monocrotaline N-oxide with human fecal suspension. 1
monocrotaline; 2-6 n-butanol extracts after incubation for 0, 1, 3,
4.5, 6 h; 7 monocrotaline N-oxide; 8-12 the corresponding H2O
layers after n-butanol extraction; (D) Control incubation of
monocrotaline N-oxide in GAM broth without feces. 1-7 n-butanol
extracts after incubation for 0, 3, 7, 11, 19, 24, 31 h; 8-14 the
corresponding H2O layers after n-butanol extraction; 15
monocrotaline N-oxide; 16 monocrotaline; (E) Incubation of isoline
or monocrotaline with human fecal suspension. 1-5 control
incubates (without PAs) for 0, 25, 50, 75, 120 h in sequence; 6-10
monocrotaline incubates for 0, 25, 50, 75, 120 h in sequence; 11-
15 isoline incubates for 0, 25, 50, 75, 120 h in sequence
This research grants from the Natural Science Foundation
of China (Nos. 31270401 and 39825129) and Wuhan University
Start-up Funding for the Introduced Talents to Dr. Jun Tang
(306276214) are greatly acknowledged.
REFERENCES
1. J. Tang and M. Hattori, Acta Pharm. Sin., 46, 762 (2011).
2. R.J. Huxtable, Pharmacol. Ther., 47, 371 (1990).
3. J. Tang, T. Akao, N. Nakamura, Z.T. Wang, K. Takagawa, M. Sasahara
and M. Hattori, Drug Metab. Dispos., 35, 1832 (2007).
4. J. Tang, M. Zhang, Z.T. Wang, T. Akao, N. Nakamura and M. Hattori,
J. Anal. Toxicol., 28, 11 (2004).
5. L. Jiang, M. Cheng, Q. Gao and J. Tang, China Pharm., 13, 1743 (2010).
6. E. Roeder, Pharmazie, 55, 711 (2000).
7. P.P. Fu, Q.S. Xia, G. Lin and M.W. Chou, Drug Metab. Rev., 36, 1
(2004).
8. Y.P. Wang, J. Yan, P.P. Fu and M.W. Chou, Toxicol. Lett., 155, 411
(2005).
9. G. Powis, M.M. Ames and J.S. Kovach, Cancer Res., 39, 3564 (1979).
10. P.S. Chu, M.W. Lame and H.J. Segall, Arch. Toxicol., 67, 39 (1993).
11. J. Tang, M. Zhang, Z.T. Wang, T. Akao, N. Nakamura, M. Hattori and
G.J. Xu, Chin. J. Nat. Med., 1, 21 (2003).
Conclusion
12. J. Tang, M. Zhang, Z.T. Wang, T. Akao, N. Nakamura and M. Hattori,
J. Chin. Pharm. Univ., 34, 499 (2003).
In this study, the biotransformation of two hepatotoxic
PAs and their N-oxides were conducted using a convenient
and non-invasive in vitro model. Moreover, a specific TLC
method employing Ehrlich reagent was developed for the
examination of PAs and their metabolites. The results revealed
that two PA N-oxides but not their parent alkaloids could be
rapidly and extensively transformed by human IB, which may
be partially responsible for the cyclic conversion between each
PA and its N-oxide and play an essential role in the toxic out-
comes of the PAs or PA-containing Chinese medicinal herbs
in Chinese medical practice.
13. L.Q. Wang, M.R. Meselhy, Y. Li, G.W. Qin and M. Hattori, Chem.
Pharm. Bull., 48, 1606 (2000).
14. L. Susag, M. Parvez, S. Mathenge and M.H. Benn, Phytochemistry,
54, 933 (2000).
15. C.G. Logie, M.R. Grue and J.R. Liddell, Phytochemistry, 37, 43 (1994).
16. S.R. Dueker, M.W. Lame, D. Morin, D.W. Wilson and H.J. Segall,
Drug Metab. Dispos., 20, 275 (1992).
17. M.W. Chou, Y.P. Wang, J. Yan, Y.C. Yang, R.D. Beger, L.D. Williams,
D.R. Doerge and P.P. Fu, Toxicol. Lett., 145, 239 (2003).