The large-scale isolation of deoxypodophyllotoxin from rhizomes of Anthriscus sylvestris followed by its bioconversion into 5- methoxypodophyllotoxin β-D-glucoside by cell cultures of Linum flavum

Article abstract of DOI:10.1021/np960748o

Dried rhizomes of Anthriscus sylvestris contained 0.39% deoxypodophyllotoxin (1). An isolation procedure with an extraction efficiency of 51.2% was developed for this commercially unavailable lignan. The isolated 1 was subsequently complexed with dimethyl

Full text of DOI:10.1021/np960748o

J . Nat. Prod. 1997, 60, 401-403
401
Th e La r ge-Sca le Isola tion of Deoxyp od op h yllotoxin fr om Rh izom es of
An th r iscu s sylvestr is F ollow ed by Its Biocon ver sion in to
5-Meth oxyp od op h yllotoxin â-D-Glu cosid e by Cell Cu ltu r es of Lin u m fla vu m
Wim Van Uden,* J eroen A. Bos, Gesiena M. Boeke, Herman J . Woerdenbag, and Niesko Pras
Department of Pharmaceutical Biology, University Centre for Pharmacy, Groningen Institute for Drug Studies (GIDS),
University of Groningen, A. Deusinglaan 1, NL-9713 AV Groningen, The Netherlands
Received December 4, 1996X
Dried rhizomes of Anthriscus sylvestris contained 0.39% deoxypodophyllotoxin (1). An isolation
procedure with an extraction efficiency of 51.2% was developed for this commercially unavailable
lignan. The isolated 1 was subsequently complexed with dimethyl-â-cyclodextrin and fed to
cell suspension cultures of Linum flavum by which it was bioconverted into 5-methoxypodo-
phyllotoxin-â-D-glucoside (2). After 7 days the cells contained 4.41% of this product on a dry-
weight basis. An isolation procedure for compound 2 was developed, with an extraction
efficiency of 83.4%.
Podophyllotoxin (3) is a naturally occurring lignan,
that is extracted from the rhizomes of Podophyllum
peltatum and Podophyllum hexandrum (Berberidaceae).
It serves as a starting compound for the preparation of
the semisynthetic cytostatics etoposide (VP-16-213) and
teniposide (VM-26).1 The supply of P. hexandrum
rhizomes, which contain ca. 4% of compound 3 on a dry-
weight basis, has become limited due to both intensive
collection and lack of cultivation.2 Therefore, the pro-
duction of 3 and related lignans by means of biotech-
nological procedures would be an interesting alterna-
tive. In the biosynthetic pathway for aryltetralin
lignans as proposed by Broomhead et al.,3 the direct
precursor of 3 is deoxypodophyllotoxin (1).
Recently, 1 was chosen as a substrate for bioconver-
sion experiments with cell suspensions of P. hexandrum,
which normally accumulate lignan 3. Because 1 is
poorly water soluble, it was added to the cell cultures
as a dimethyl-â-cyclodextrin complex. In addition,
similar experiments were performed with cell suspen-
sions of Linum flavum (Linaceae; yellow flax), which
normally accumulate 5-methoxypodophyllotoxin (4).4
Compound 4 has a cytotoxicity that is comparable to
that of 3, so it may also be of interest as a starting
compound for new or improved cytostatics.5 The P.
hexandrum cell suspensions converted 1 into 3, while
in the L. flavum cultures 5-methoxypodophyllotoxin â-D-
glucoside (2) was formed. The contents as measured
after bioconversion were high, respectively: 2.9% and
2.4% on a dry-weight basis.
Because lignan 1 is commercially unavailable, it must
be isolated from a suitable natural source for use as a
substrate in bioconversion studies. Compound 1 occurs
in the rhizomes of Anthriscus sylvestris (Apiaceae; wild
chervil) a common perennial indigenous to The Neth-
erlands.6,7
In this paper, a large-scale isolation procedure of
compound 1 from the rhizomes of A. sylvestris is
presented. In addition, its bioconversion into the gluco-
side 2 by cell cultures of L. flavum and the subsequent
isolation of this product is described.
The A. sylvestris rhizomes lost 71% H2O upon drying
and contained 0.39 ( 0.02% of 1 on a dry-weight basis
(n ) 10), corresponding to 6.83 g in the total amount of
1.75 kg of plant material. After Soxhlet extraction,
86.1% of 1 originally present in the plant material was
extracted. In the extracted rhizomes 1 could not be
detected anymore, thus 13.9% of 1 was lost during the
extraction procedure, probably due to decomposition.
The first column chromatographic purification of the
extract caused no loss of 1, but (colored) impurities were
removed. During the second column chromatographic
purification 4.8% of 1 was lost. The HPLC chromato-
gram of the collected fractions now showed two peaks
one of compound 1 and one of an unknown compound.
The latter was removed on an Amberlite XAD-2 column.
The peak ratio of compound 1:unknown compound
increased from 1.28 to 1.75, while the loss of 1 was only
1.5%. Finally, crystallization yielded 3.5 g of 1, corre-
sponding to an extraction efficiency of 51.2%.
On the basis of HPLC analysis, the isolated 1 had a
purity of ca. 98%. TLC showed a spot with the same
color and with the same Rf value (0.71) as the reference
compound 1 (prepared by chemical reduction of 3). The
melting point of 1 is in agreement with earlier reported
values of 167.4-168.3 °C.11 The IR spectrum as well
* To whom correspondence should be addressed. Phone: ++31-50-
3632769. FAX: ++31-50-3637570. E-mail: w.van.uden@farm.rug.nl.
X Abstract published in Advance ACS Abstracts, April 1, 1997.
S0163-3864(96)00748-3 CCC: $14.00
© 1997 American Chemical Society and American Society of Pharmacognogy

402 J ournal of Natural Products, 1997, Vol. 60, No. 4
Notes
as the 13C-NMR spectrum corresponded with those of
the reference 1.
unique combination of an unlimitedly available sub-
strate and cell cultures with a high bioconversion
capacity offer possibilities for the biotechnological pro-
duction of the cytotoxic lignans 3 and 4.
The presence of cyclodextrin complexed 1 had no effect
on the growth of the L. flavum cell cultures in terms of
packed cell volume, dry weight, fresh weight, and
conductance. During the past 5 years, the use of
cyclodextrins in order to solubilize apolar substrates in
aqueous environments has proven to create smooth
conditions for various bioconversions carried out in our
laboratory.11
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. Compound 1
(prepared at the Department of Organic and Molecular
Inorganic Chemistry, University of Groningen, The
Netherlands), compounds 2 (kindly provided by TNO-
Zeist, The Netherlands) and 4 (isolated from cell cul-
tures of L. flavum8) were analyzed by HPLC on a
Lichrosorb 7RP-18 column (100 × 3 mm i.d.; Chrompack,
Middelburg, The Netherlands) with MeOH-H2O (45:
55; v/v) as the mobile phase at a flow rate of 1.0 mL
min-1 and detection at 290 nm.9 For the determination
of 1 in the dried rhizomes of Anthriscus sylvestris, 50
mg was extracted with 4.0 mL MeOH for 1 h followed
by centrifugation (5 min, 1500 g). For the determination
of 2 in Linum flavum cells, 20 mg of dried material was
extracted by ultrasonication with 2 mL of various
mixtures of MeOH-H2O for 1 h followed by centrifuga-
tion (5 min, 1500 g). For the hydrolysis of 2 into its
aglucone (4), 1.0 mL of the above mentioned extracts of
L. flavum cells, was evaporated to dryness and redis-
solved in 0.10 mL MeOH. Then 2 mL of a 2.5% (m/v)
solution of â-glucosidase (Sigma, Chemical Co.) in
phosphate buffer (0.1 M, pH ) 5.0) was added, and the
mixture was incubated at 37 °C. Samples of 0.5 mL
were taken after 2, 5, 6, and 72 h. After addition of 2.0
mL MeOH they were vortexed and centrifuged (5 min,
1500 g). TLC was performed using Si gel F254 plates
(Merck, cat. no. 5715). The mobile phase was CHCl3-
MeOH (100:4; v/v). The spots were visualized after a
10 s bath in H2SO4-MeOH (1:20; v/v) and subsequent
heating at 110 °C for 10 min. IR spectra (KBr disks)
were recorded on a ATI Mattson FTIR infrared spectro-
photometer. 13C-NMR spectra (Attached Proton Test)
were recorded on a Varian VXR-500 (125 MHz) system
with CDCl3 as the solvent. The melting points were
determined using an Electrothermal IA 9100 melting
point apparatus an are uncorrected.
Extr a ction a n d Isola tion of Deoxyp od op h yllo-
toxin (1). Dried rhizomes of A. sylvestris (1.75 kg) were
macerated overnight in 18 L of MeOH followed by
Soxhlet extraction for 14 h, and the solvent was removed
under reduced pressure. The residue was subjected to
column chromatography (20 × 11 cm i.d.) over Si gel
60 (Fluka Chemika cat. no. 60741; 1045 g) with n-hex-
ane-Me2CO (2:1; v/v) as the eluent. Fractions contain-
ing 1 were pooled and evaporated to dryness using a
rotary evaporator. Columm chromatography (120 × 10
cm i.d.; Si gel 60, 3711 g) was repeated with the same
eluent. Pooled fractions containing 1 were dried under
reduced pressure and chromatographed (70 × 5 cm i.d.)
over Amberlite XAD-2 (Fluka Chemika cat. no. 06443;
678 g) with MeOH as the eluent. Pooled fractions
containing 1 were evaporated to dryness using a rotary
evaporator and redissolved in 100 mL MeOH of 60 °C
and the volume was slowly reduced to about 50 mL
under stirring and a gentle stream of air. After stirring
overnight, crystals of compound 1 were harvested by
filtration on a Bu¨chner-funnel, washed with a small
quantity of ice-cold MeOH and dried at room temper-
ature. The yield was 3.5 g of 1 with a purity of 98%
based on the peak heights in the HPLC chromatogram.
Recently, compound 4 and its â-D-glucoside (2) were
produced in cell cultures of L. flavum by bioconversion
of cyclodextrin-complexed 1 (1.4 mM).4 After 7 days
2.38% of combined 2 and 4 on a dry-weight basis was
found, corresponding with a production rate of 35.6 mg
L-1 day-1. On the basis of these results, we harvested
the plant cells in the present study after 7 days of
incubation with the substrate. The cells contained
4.41% 2 on a dry-weight basis, corresponding with a
production rate of 71.2 mg L-1 day-1 and a bioconver-
sion percentage of 75.3%. This is a very high production
rate, taking the structural complexity of the lignan into
account.13 The bioconversion of compound 1 into the
glucoside 2 comprises at least four steps: two hydroxyl-
ations, a methylation, and a glucosylation.
The percentage of H2O in the extraction fluid has been
described to determine whether 2 is extracted as the
glucoside itself or as the corresponding aglucone and
can be explained by the presence of a â-glucosidase in
the cells.14 In the present study cells were extracted
with different mixtures of MeOH-H2O. Compound 2
was completely extracted as its glucoside with a MeOH
percentage of 70% or higher. Lower concentrations of
MeOH increased the amount of aglucone. The total
amount of extracted 2 either as aglucone or as glucoside
was maximal between 20% and 80% MeOH. On the
basis of these results a mixture of 80% MeOH and 20%
H2O was used in order to isolate the lignan 4 completely
in its glucosidic form. It was proven that in L. flavum
cell cultures 1 is bioconverted into 2 and not into its
corresponding aglucone. After purification of the extract
by column chromatography on Si gel (no loss), the
combined fractions containing the glucoside were ex-
tracted with CH2Cl2. Repeating the extraction 12 times
caused a loss of only 1.5%. HPLC analysis showed that
the combined fractions did not contain only pure glu-
coside 2 but that also an unknown impurity was
present. Final purification was achieved by column
chromatography over Si gel with PhMe-Et2O-MeOH
as the eluent. This step resulted in 15.3% loss of 2. The
yield of 2 with a purity of 99% based on HPLC analysis
was 2.0 g, equivalent to an extraction efficiency of
83.4%. The identity of the glucoside was confirmed by
adding â-glucosidase to a solution of 2, which resulted
in the formation of 4. Finally, 13C-NMR data of 2 were
in agreement with those published previously for com-
pound 4,8,15 and six additional peaks corresponding with
the carbon atoms of the attached glucose moiety were
present.
In conclusion, a procedure was developed to obtain
large quantities of compound 1 from Anthriscus sylves-
tris. This lignan was shown to be a suitable substrate
for bioconversion into 2 by cell cultures of L. flavum at
a high production rate. The bioconversion product was
successfully isolated using the procedure described. The

Notes
Biocon ver sion of Deoxyp od op h yllotoxin (1) in to
J ournal of Natural Products, 1997, Vol. 60, No. 4 403
1036, 996 cm-1 13C NMR (CDCl3) δ 174.22 (C-12),
;
5-Meth oxyp od op h yllotoxin -â-D-glu cosid e (2). Cell
suspension cultures of L. flavum were obtained and
grown as described previously.10 For bioconversion, 2.39
g (6 mmol) of the isolated 1 was dissolved in 105 mL
MeOH-n-BuOH (1/4; v/v). To three sterilized 1-L
bottles, 35 mL of this solution was added aseptically
through a 0.2-µL filter. The solvent was evaporated in
a laminar air-flow cabinet. Culture media were supple-
mented with 2,6-dimethyl-â-cyclodextrin (Avebe, Veen-
dam, The Netherlands; 4 mM) and autoclaved (121 °C,
20 min). Of this medium 1 L was added to the bottles
containing 1, and they were shaken until 1 had com-
pletely dissolved. In 500-mL flasks, 200 mL of medium
with 1 as a cyclodextrin complex was inoculated with
125 mL of a 2-week-old cell suspension of L. flavum,
yielding a final concentration of 1 of 1.23 mM. The cells
were harvested after 7 days by filtration and frozen at
-20 °C followed by lyophilization.
Isola tion a n d P u r ifica tion of 5-Meth oxyp od o-
p h yllotoxin â-D-glu cosid e (2). Dried L. flavum cells
(55.2 g) were extracted with 2 L of MeOH 80% (v/v) by
ultrasonication for 1.5 h. After filtration and squeezing,
the cell material was extracted again with 700 mL of
80% (v/v) MeOH. The solvent was removed by rotary
evaporation yielding 22 g of raw extract, which was
further extracted by ultrasonication with 300 mL H2O
for 40 min. The suspension obtained was chromato-
graphed on a column (20 × 11 cm i.d.), with 850 g Si
gel 60; H2O was used as the eluent at 10 mL/min.
Fractions containing 2 were pooled, concentrated to a
volume of 600 mL under reduced pressure, and ex-
tracted 12 × with 300-mL portions of CH2Cl2. After
each extraction 2 was determined in the H2O phase. The
CH2Cl2 fractions yielded 2.6 g residue that was sub-
jected to column chromatography (97 × 1.8 cm i.d.) over
Si gel 60 (95 g) with PhMe-Et2O-MeOH (10:10:7; v/v/
v) as the eluent. Fractions containing 2 were pooled
and dried under reduced pressure. The residue was
redissolved in 50 mL H2O, frozen at -20 °C and
lyophilized. The yield was 1.0 g of 2 with a purity of
99% on the basis of the peak heights in the HPLC
chromatogram.
152.36 (C-3′ + C-5′), 149.71 (C-7), 141.89 (C-5), 136.72
(C-6), 134.74 (C-1′), 134.50 (C-4′), 132.64 (C-9), 122. 18
(C-10), 107.98 (C-2′ + C-6′), 104.75 (C-8), 101.66 (C-13),
71.97 (C-11), 69.16 (C-4), 60.60 (C-4′ - OCH3), 60.01
(C-5 - OCH3), 56.30 (C-3′ - OCH3 + C-5′ - OCH3),
45.79 (C-2), 44.19 (C-1), 38.66 (C-3) glucoside: ad-
ditional δs at 76.30, 75.65, 74.34, 72.93, 56.15 (CH’s),
and 61.24 (CH2).
P la n t Ma ter ia l. Plants of A. sylvestris Hoffm. (Apia-
ceae) were harvested on J une 2, 1993, from a road side
at Lage Meeden, The Netherlands. A voucher specimen
is deposited in our institute. The rhizomes were cut
from the plants, and the clay was removed by washing
with H2O. The rhizomes were freeze dried, cut into
pieces, ground, and sieved (1 mm).
Ack n ow led gm en t. We thank Mr. J . F. Bracht
Waker for performing preliminary experiments concern-
ing the isolation of deoxypodophyllotoxin from the
rhizomes of Anthriscus sylvestris. We thank Dr. A. J .
R. L. Hulst and Mr. W. H. Kruizinga from the Depart-
ment of Organic and Molecular Inorganic Chemistry,
University of Groningen, The Netherlands, for recording
the NMR spectra. We also thank Mr. G. J . Rozendal
for the help with the final purification step of the
glucoside.
Refer en ces a n d Notes
(1) Sta¨helin, H. F.; Von Wartburg, A. Prog. Drug Res. 1989, 33, 169-
266.
(2) Gupta, S.; Sethi, K. L. In Conservation of Tropical Plant
Resources. Botanical Survey of India. J ain, S. K., Mehra, K. L.,
Eds.; Howrah: New Delhi, pp 101-107.
(3) Broomhead, A. J .; Rahman, M. A.; Dewick, P. M. J ackson, D.
E.; Lucas, J . A. Phytochemistry 1991, 30, 1489-1492.
(4) Van Uden, W.; Bouma, A. S.; Bracht Waker, J . F.; Middel, O.;
Wichers, H. J ., De Waard, P.; Woerdenbag, H. J .; Kellogg, R.
M.; Pras, N. Plant Cell Tiss. Org. Cult. 1995, 42, 73-79.
(5) Middel, O.; Woerdenbag, H. J .; Van Uden, W.; Van Oeveren, A.;
J ansen, J . F. G. A.; Feringa, B. L.; Konings, A. W. T.; Pras, N.;
Kellogg, R. M. J . Med. Chem. 1995, 38, 2112-2118.
(6) Noguchi, K.; Kawanami, M. J . Pharm. Soc. J apan 1940, 60,
629-636; Chem. Abstr. 1953, 47, 6386.
(7) Weeda, E. J .; Westra, R.; Westra, C.; Westra, T. Nederlandse
Oecologische Flora: Wilde Planten en hun Relaties 2. IVN:
Amsterdam, 1987; pp 252-254.
(8) Van Uden, W.; Homan, B.; Woerdenbag, H. J .; Pras, N.;
Malingre´, T. M.; Wichers, H. J .; Harkes, M. J . Nat. Prod. 1992,
55, 102-110.
Deoxyp od op h yllot oxin (1): white crystals, mp
166.7-168.2 °C (MeOH); IR νmax (KBr) 2937, 2842, 1765
(CdO), 1587, 1504, 1418, 1331, 1225, 1123, 1034, 941,
(9) Van Uden, W.; Pras, N.; Visser, J . F.; Malingre´, T. M. Plant Cell
Rep. 1989, 8, 165-168.
(10) Van Uden, W.; Pras, N.; Batterman, S., Visser, J . F.; Malingre´,
T. M. Planta 1991, 183, 25-30.
857, 768 cm-1 13C NMR (CDCl3) δ 174.8 (C-12), 152.3
;
(11) Hartwell, J . L.; Schrecker, A. W. (1958) In Progress in the
Chemistry of Organic Natural Products; Zechmeister, L., Ed.;
Springer-Verlag: Vienna, 1958; Vol. 15, pp 83-166.
(12) Van Uden, W.; Woerdenbag, H. J .; Pras, N. Plant Cell Tiss. Org.
Cult. 1994, 38, 103-113.
(C-3′ + C-5′), 146.9 (C-7), 146.6 (C-5), 136.8 (C-6), 136.1
(C-10), 130.5 (C-1′), 128.1 (C-4′), 110.3 (C-9), 108.3 (C-
8), 108.1 (C-2′ + C-6′), 101.0 (C-13), 71.9 (C-11), 60.6
(C-4′ - OCH3), 56.1 (C-3′ - OCH3 + C-5′ - OCH3), 47.3
(C-2), 43.6 (C-1), 33.0 (C-4), 32.6 (C-3).
5-Met h oxyp od op h yllot oxin â-D-glu cosid e (2):
white crystals, mp 155-165 °C (H2O) with efferves-
cence; IR νmax (KBr) 3425 (OH), 2938, 2902, 2839, 1774
(CdO), 1618, 1589, 1507, 1476, 1420, 1252, 1125, 1071,
(13) Pras, N.; Woerdenbag, H. J .; Van Uden, W. Agbiotech News
Information 1995, 7, 231N-243N.
(14) Berlin, J .; Bedorf, N.; Mollenschott, C.; Wray, V.; Sasse, F.; Ho¨fle,
G. Planta Med. 1988, 54, 204-206.
(15) Wichers, H. J .; Harkes, M. P.; Arroo, R. J . Plant Cell Tiss. Org.
Cult. 1990, 23, 93-100.
NP960748O