Dirigent-mediated podophyllotoxin biosynthesis in Linum flavum and Podophyllum peltatum

Article abstract of DOI:10.1016/S0031-9422(00)00242-9

Given the importance of the antitumor/antiviral lignans, podophyllotoxin and 5-methoxypodophyllotoxin, as biotechnological targets, their biosynthetic pathways were investigated in Podophyllum peltatum and Linum flavum. Entry into their pathways was established to occur via dirigent mediated coupling of E-coniferyl alcohol to afford (+)-pinoresinol; the encoding gene was cloned and the recombinant protein subsequently obtained. Radiolabeled substrate studies using partially purified enzyme preparations next revealed (+)-pinoresinol was enantiospecifically converted sequentially into (+)-lariciresinol and (-)-secoisolariciresinol via the action of an NADPH-dependent bifunctional pinoresinol/lariciresinol reductase. The resulting (-)-secoisolariciresinol was enantiospecifically dehydrogenated into (-)-matairesinol, as evidenced through the conversion of both radio- and stable isotopically labeled secoisolariciresinol into matairesinol, this being catalyzed by the NAD-dependent secoisolariciresinol dehydrogenase. (-)-Matairesinol was further hydroxylated to afford 7'-hydroxymatairesinol, this being efficiently metabolized into 5-methoxypodophyllotoxin. Thus much of the overall biosynthetic pathway to podophyllotoxin has been established, that is, from the dirigent mediated coupling of E-coniferyl alcohol to the subsequent conversions leading to 7'-hydroxymatairesinol. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0031-9422(00)00242-9

Phytochemistry 55 (2000) 537±549
www.elsevier.com/locate/phytochem
Dirigent-mediated podophyllotoxin biosynthesis in Linum ¯avum
and Podophyllum peltatum
Zhi-Qiang Xia, Michael A. Costa, John Proctor, Laurence B. Davin, Norman G. Lewis *
Institute of Biological Chemistry, Washington State University, Pullman, PO Box 646340, WA 99164-6340, USA
Received 9 May 2000
Abstract
Given the importance of the antitumor/antiviral lignans, podophyllotoxin and 5-methoxypodophyllotoxin, as biotechnological
targets, their biosynthetic pathways were investigated in Podophyllum peltatum and Linum ¯avum. Entry into their pathways was
established to occur via dirigent mediated coupling of E-coniferyl alcohol to a€ord (+)-pinoresinol; the encoding gene was cloned
and the recombinant protein subsequently obtained. Radiolabeled substrate studies using partially puri®ed enzyme preparations
next revealed (+)-pinoresinol was enantiospeci®cally converted sequentially into (+)-lariciresinol and (� )-secoisolariciresinol via
the action of an NADPH-dependent bifunctional pinoresinol/lariciresinol reductase. The resulting (� )-secoisolariciresinol was
enantiospeci®cally dehydrogenated into (� )-matairesinol, as evidenced through the conversion of both radio- and stable iso-
topically labeled secoisolariciresinol into matairesinol, this being catalyzed by the NAD-dependent secoisolariciresinol dehy-
0
drogenase. (� )-Matairesinol was further hydroxylated to a€ord 7 -hydroxymatairesinol, this being eciently metabolized into 5-
methoxypodophyllotoxin. Thus much of the overall biosynthetic pathway to podophyllotoxin has been established, that is, from the
0
dirigent mediated coupling of E-coniferyl alcohol to the subsequent conversions leading to 7 -hydroxymatairesinol. # 2000 Elsevier
Science Ltd. All rights reserved.
0
Keywords: Podophyllum peltatum; Linum ¯avum; Berberidaceae; Linaceae; Dirigents; 7 -Hydroxymatairesinol; 5-Methoxypodophyllotoxin;
Podophyllotoxin; Pinoresinol/lariciresinol reductase; Secoisolariciresinol dehydrogenase
1
. Introduction
etoposide (2a), etopophos (2b) and teniposide (3), semi-
synthetic derivatives of podophyllotoxin (1) which do
not display the same high levels of cytotoxicity, are
widely used clinically for treatment of Hodgkin's and
non-Hodgkin's lymphoma, testicular/small cell lung
cancers and acute leukemia (O'Dwyer et al., 1985;
Williams et al., 1987; Young, 1992; Schacter, 1996;
Canel et al., 2000).
Because of the limited supply of Podophyllum rhizomes,
due to their intensive collection in the wild, there is
considerable interest in both de®ning and biotechnolo-
gically exploiting the podophyllotoxin (1) biosynthetic
pathway, thereby making this compound more generally
available (Lewis and Davin, 1999). This need is driven
in part because studies, using both cell suspension and
root tissue cultures of Podophyllum and Linum species,
have had limited success thus far in attaining sig-
ni®cantly elevated levels of either podophyllotoxin (1)
or its 5-methoxy analogue (4) (van Uden et al., 1990a;
Podophyllotoxin (1) (Scheme 1) is a very important
antitumor and antiviral agent isolated from Podo-
phyllum species. Its pharmacological usage dates back
many centuries, when may apple (P. peltatum) alcoholic
extracts, obtained from rhizomes and roots, were
employed ®rst as a poison and later, in smaller doses,
for treatment of various pathological conditions (Ayres
and Loike, 1990). It was subsequently demonstrated
that the cytotoxic e€ect of these extracts was due to the
lignan, podophyllotoxin (1). However, direct applica-
tion/administration of podophyllotoxin (1) was accom-
panied with severe toxicity, which precluded its direct
use in cancer treatment. Podophyllotoxin (1) has since
found direct medicinal usage in an e€ective treatment of
venereal warts (Beutner and von Krogh, 1990). Today,
*
Corresponding author. Tel.: +1-509-335-8382; fax: +1-509-335-
206.
E-mail address: lewisn@wsu.edu (N.G. Lewis).
1990b). Accordingly, we are deciphering the biosyn-
thetic pathway to the medicinally important lignan,
8
0
031-9422/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0031-9422(00)00242-9

5
38
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
podophyllotoxin (1), as a ®rst step towards the bio-
technological exploitation of its biochemical pathway.
This paper describes recent progress towards meeting
these goals.
1996) and secoisolariciresinol dehydrogenase (Umezawa
et al., 1991; Xia et al., 2000) in downstream metabolism.
All of these proteins and enzymes were puri®ed from F.
intermedia with the corresponding genes cloned and
fully functional recombinant proteins expressed.
Accordingly, with both P. peltatum and L. ¯avum plants
available for study, where the former accumulates
podophyllotoxin (1) and the latter, its 5-methoxy ana-
logue (4), we wished to establish their biochemical
pathway(s). Of particular interest was whether matair-
esinol (5) was an intermediate and, if so, the steps
involved in the biochemical pathway to matairesinol (5)
from E-coniferyl alcohol (7).
We sought ®rst to establish whether the biochemical
pathway to matairesinol (5) in Linum ¯avum, which
accumulates 5-methoxypodophyllotoxin (4), was as for
Forsythia sp. That is, it was established that, following
dirigent protein mediated (+)-pinoresinol (6a) forma-
tion (Davin et al., 1997), this then undergoes sequential
enantiospeci®c reduction to give (+)-lariciresinol (8a)
and (� )-secoisolariciresinol (9b) (Chu et al., 1993;
2. Results and discussion
Previous work had suggested the possibility of a bio-
chemical precursor relationship between matairesinol
5) and podophyllotoxin (1) (Broomhead et al., 1991)
(
although, based on incorporation levels for Podo-
phyllum species, it was not possible to conclude that this
was a direct intermediate in the pathway. A biosynthetic
pathway to matairesinol (5) had, however, been rigor-
ously established in Forsythia species (Scheme 2): this
occurred through the action of the dirigent protein
a€ording (+)-pinoresinol (6a) from E-coniferyl alcohol
(
7) (Davin et al., 1997; Gang et al., 1999a), as well as
that of pinoresinol/lariciresinol reductase (Chu et al.,
993; Katayama et al., 1993; Dinkova-Kostova et al.,
1
Scheme 1. Podophyllotoxin (1), its derivatives (2a, 2b and 3) and an analogue (4).
Scheme 2. Biosynthetic pathway from E-coniferyl alcohol (7) to matairesinol (5) in Forsythia intermedia. Note: compounds designated with an ``a''
indicate the (+)-antipode, whereas those with a ``b'' depict the (� )-enantiomer. [In each case, only one enantiomeric form is illustrated.]

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
539
Katayama et al., 1993; Dinkova-Kostova et al., 1996),
followed by enantiospeci®c dehydrogenation to a€ord
It was next instructive to ascertain whether secoiso-
lariciresinol (9) would be converted into matairesinol (5)
when incubated in the presence of NAD with L. ¯avum
cell-free extracts. In order to investigate this question,
(
� )-matairesinol (5b) (Umezawa et al., 1991). Thus, cell-
free extracts from L. ¯avum roots were ®rst incubated
with racemic (Æ)-pinoresinols (6a/6b) in the presence of
2
both (Æ)-[Ar- H]-secoisolariciresinols (9a/9b) and (Æ)-
3
0
3
[
4R- H]-NADPH. Following 2 h incubation, the resulting
[9,9 - H ]-secoisolariciresinols (9a/9b) were synthesized
2
assay mixture (see Experimental) was extracted with
EtOAc containing both racemic (Æ)-lariciresinols (8a/8b)
and (Æ)-secoisolariciresinols (9a/9b) as radiochemical
carriers. The resulting EtOAc solubles were then applied
to a C18 reversed phase HPLC column eluted with
and individually employed as potential substrates. Thus,
2
0.55 mmol (Æ)-[Ar- H]-secoisolariciresinols (9a/9b) were
incubated in the presence of 40 mM NAD, with a par-
tially puri®ed enzyme preparation from L. ¯avum roots,
obtained via consecutive DEAE cellulose and ADP-
Sepharose column chromatographic steps. After 2 h,
the resulting assay mixture was extracted with EtOAc,
with the EtOAc solubles combined and applied to a
MeOH±3%HOAc in H O (30:70), with fractions corre-
2
sponding to lariciresinol (8) and secoisolariciresinol (9)
individually collected and evaporated to dryness.
Reconstitution of each sample in EtOH±hexanes (1:1),
followed by chiral HPLC analysis (Figs. 1A±D), revealed
that radioactivity was only associated with (+)-lariciresinol
C18 reversed phase HPLC column eluted with CH CN±
3
3% HOAc in H O. The fraction corresponding to
2
matairesinol (5) was collected and subjected to electron
impact mass spectroscopic analysis (see Fig. 2). For
comparative purposes, Fig. 2A shows the mass spec-
trum of unlabeled matairesinol (5), whose molecular ion
(
1
8a) (Fig. 1C) and (� )-secoisolariciresinol (9b) (Fig.
D). As noted, no radioactivity was detectable in either
(
� )-lariciresinol (8b) or (+)-secoisolariciresinol (9a),
+
even with unlabeled racemic lignans having been added
as radiochemical carriers (Figs. 1A and B). Moreover,
the sequential reductive steps abstracted only the
[M ] was clearly observed at m/z 358, together with its
benzylic base peak ion at m/z 137 (Umezawa et al.,
3
[
4R- H]-NADPH hydride at the 4R-position and not
that from the corresponding 4S-position (data not
shown). Thus the biosynthetic pathway from E-con-
iferyl alcohol (7) to secoisolariciresinol (9) in L. ¯avum
root tissue was as for F. intermedia. This result could
not, however, have been predicted a priori since lignans
are found in di€erent plant species with di€ering
degrees of enantiomeric purity (Lewis and Davin, 1999).
Fig. 1. A and B: Chiral column HPLC separations of (A) synthetic
(
Æ)-lariciresinols (8a/8b) and (B) (Æ)-secoisolariciresinols (9a/9b) on
Chiralcel OD and OC columns, respectively; these racemates were
used as radiochemical carriers. C and D: Chiral HPLC analyses with
radiochemical detection of enzymatic products, (C) lariciresinol (8)
and (D) secoisolariciresinol (9) following incubation of a crude pinor-
esinol/lariciresinol reductase preparation from L. ¯avum with (Æ)-
3
pinoresinols (6a/6b) (0.4 mM) in the presence of [4R- H]-NADPH (0.8
Fig. 2. Mass spectral fragmentation pattern (EI mode) of (A) natural
2
mM). Elution conditions: Chiralcel OD: hexanes±EtOH (70:30); ¯ow
�
abundance matairesinol (5) and (B) [Ar- H]-matairesinol (5) obtained
2
rate, 0.5 ml min 1. Chiralcel OC: hexanes±EtOH (20:80); ¯ow rate,
following incubation of [Ar- H]-secoisolariciresinol (9) with a partially
puri®ed L. ¯avum secoisolariciresinol dehydrogenase preparation.
�
0
.2 ml min 1.

5
40
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
1991). The mass spectrum of the enzymatic product
obtained following incubation with [Ar- H]-secoisolar-
2
2
iciresinols (9a/9b) con®rmed the product to be [Ar- H]-
matairesinol (5) (Fig. 2B). That is, there was a molecular
+
ion cluster centered at m/z 360 [M +2] corresponding
to deuterated matairesinol (5), with additional
con®rmation that the deuterium was in the aromatic
ring being the base peak ion fragment centered at m/z
+
2
1
38 [Ar +1]. Thus, [Ar- H]-secoisolariciresinol (9)
2
had been intactly converted into [Ar- H]-matairesinol
5).
To determine the enantiospeci®city of this conversion,
(
the partially puri®ed secoisolariciresinol dehydrogenase
preparation was next incubated for 2 h with 2.8 mM (Æ)-
0
9,9 - H ]-secoisolariciresinols (9a/9b) in the presence of
2
3
[
excess NAD (40 mM). Following incubation, the corre-
sponding radiolabeled matairesinol (5) was isolated
using C18 reversed phase HPLC as before, where it was
0
3
subsequently demonstrated that only (� )-[9,9 - H ]-seco-
2
3
isolariciresinol (9b) was converted into (� )-[9- H ]-matai-
2
Fig. 3. C18 reversed phase HPLC chromatogram of (A) a MeOH
extract obtained after incubation of matairesinol (5) with L. ¯avum
resinol (5b), i.e. the corresponding (+)-enantiomer (9a) was
not utilized. The enantiospeci®city of the conversion
was demonstrated by chiral HPLC analysis: however,
since chiral column HPLC (Chiralcel OD) does not
completely separate both (+)- and (� )-antipodes of
matairesinol (5a) and (5b) (data not shown), the
matairesinol (5) that had been enzymatically produced
0
root tissue, and (B) authentic 7 -hydroxymatairesinol (16). Elution
conditions: linear gradient CH CN±3%HOAc in H O from 10:90 to
0:70 between 0 and 35 min; ¯ow rate 1 ml min 1
3
2
�
3
.
was chemically reduced with LiAlH (see Experimental)
4
to a€ord secoisolariciresinol (9). Chiral HPLC analysis
revealed that only the (� )-antipode (9b) was radio-
labeled (data not shown) and that the conversion was
enantiospeci®c. This observation was in agreement with
previous results with F. intermedia (Umezawa et al.,
1991; Xia et al., 2000). Thus, taken together, these data
established that, in the podophyllotoxin-forming species,
the biochemical pathway to (� )-matairesinol (5b) was
exactly the same as in Forsythia (see Scheme 2).
Attention was next focused on the metabolism of
matairesinol (5) in L. ¯avum root tissues, and whether it
served as a precursor of 5-methoxypodophyllotoxin (4)
or not. Thus, administration of a solution containing
0.28 mmol unlabeled matairesinol (5) to L. ¯avum root
segments was carried out for 6 h, following which the
root tissue was homogenized and then extracted with
MeOH under conditions of re¯ux. The resulting MeOH
solubles were next treated with b-glucosidase, with the
resulting EtOAc extract subjected to silica gel chromato-
graphy and C18 reversed phase HPLC, successively.
HPLC analysis revealed the presence of a new metabolite
at an elution volume of 14.3 ml (Fig. 3A), the EI mass
spectrum of which gave a molecular ion at m/z 374
+
[
M +16] (Fig. 4B). This metabolite was tentatively
0
identi®ed as 7 -hydroxymatairesinol (16), based on ana-
lysis of its mass spectral fragmentation pattern. This
interpretation was subsequently established as being
correct through comparison with an authentic sample
Fig. 4. Mass spectral fragmentation pattern (EI mode) of (A) syn-
0
0
thetic 7 -hydroxymatairesinol (16), and (B) the putative 7 -hydro-
xymatairesinol (16) product obtained after administration of
matairesinol (5) to L. ¯avum root tissue.

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
541
0
Scheme 3. Total synthesis of 7 -hydroxymatairesinol (16).
0
obtained by total synthesis (Scheme 3 and Fig. 3B). As
0
Next, we wished to establish if 7 -hydroxymatairesinol
(16) served as an intermediate in the biosynthesis of
can be seen in Scheme 3, authentic 7 -hydro-
0
3
0
xymatairesinol (16) was synthesized from an O-benzylic
derivative of vanillin (10). In this one-pot reaction, the
protected vanillin (10) was ®rst converted into its bis-
phenylthioacetal derivative (11) and then into the cor-
responding anionic form with n-BuLi. Subsequent
nucleophilic attack with 2(5H)furanone and the bromo-
derivative (12) a€orded the desired butyrolactone (13)
which was then converted into the corresponding
ketone (14) via the action of HgO/BF Et O (Ziegler and
5-methoxypodophyllotoxin (4). Thus, [7 - H]-7 -hydro-
xymatairesinol (16) was synthesized from ketone (14),
via deployment of NaB H as the reducing reagent (see
4
Scheme 3). Following deprotection, the [7 - H]-7 -
hydroxymatairesinol (16) was administered to L. ¯avum
root tissue for 6 h. The resulting methanol extract was
again treated with b-glucosidase, at which time it was
demonstrated that this precursor had been converted
3
0
3
0
3
into radiolabeled [7- H]-5-methoxypodophyllotoxin (4)
[1.3% total incorporation] (see Fig. 5).
3
2
Schwartz, 1978). Consecutive reductions, using NaBH4,
and deprotection (Pd/C, H ), respectively, a€orded
2
0
authentic 7 -hydroxymatairesinol (16).
0
The synthetic 7 -hydroxymatairesinol (16) had both
an identical retention volume of 14.3 ml (Fig. 3B) to
that of the L. ¯avum matairesinol-derived metabolite
(
Fig. 3A), as well as an identical UV spectrum (284, 228
and 202 nm). Furthermore, mass spectral comparison of
the synthetic product to that of the Linum metabolite
further con®rmed its identity (see Figs. 4A and B). The
0
authentic 7 -hydroxymatairesinol (16) (Fig. 4A) gave the
expected parent molecular ion at m/z 374 together with
two fragment ions at m/z 153 and 137 corresponding to
0
both 7 -hydroxylated benzylic and benzylic ion frag-
ments, respectively. These data thus clearly established
0
that the hydroxyl group was at position 7 . EI mass
spectral analysis of the matairesinol-derived metabolite
gave an essentially identical fragmentation pattern (Fig.
4
B), thereby con®rming its identity. Thus, matairesinol
0
(5) had been converted into 7 -hydroxymatairesinol (16)
in L. ¯avum. Note also that this compound had pre-
viously been reported in western hemlock (Tsuga het-
erophylla) (Krahmer et al., 1970), suggesting its
involvement in the biosynthesis of the lignan, a-con-
idendrin (Lewis and Davin, 1999).
Fig. 5. HPLC analysis of 5-methoxypodophyllotoxin (4) isolated from
0
3
0
L. ¯avum roots following administration of [7 - H]-7 -hydro-
xymatairesinol (16). (A) UV absorbance at 280 nm and (B) radio-
activity elution pro®les.

5
42
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
Fig. 6. GAP comparison of the PCR DNA fragment isolated from the P. peltatum rhizome cDNA library (upper sequence) and the corresponding
dirigent sequence from F. intermedia (lower sequence), indicating 80.2% similarity and 72.5% identity. Numbers to the right of the sequences indi-
cate the amino acid residue locations in the respective sequences. Boldface characters indicate the residues from which degenerate dirigent primers
were designed to obtain the P. peltatum dirigent homologue sequence. A pipe character (j) between two sequence symbols indicates an identical
residue match, a colon indicates a comparison value greater than or equal to the average positive non-identical comparison value in the scoring
matrix, and a period signi®es a comparison value greater than or equal to 1.
Fig. 7. Complete cDNA and amino acid sequence of P. peltatum dirigent homologue psd-Pp1. The numbers to the right indicate the DNA base or
amino acid sequence residue locations, respectively. A Kozak consensus DNA sequence is identi®ed by underlined letters; the translation initiation
methionine and the stop codon are in bold type; a termination consensus signal is indicated by bold italicized letters, and the polyA tail is presented
in underlined bold type. Three potential N-glycosylation sites are indicated by gray background italicized amino acid residues. Underlined bold,
italicized amino acid residue characters indicate locations where degenerate primers PS6F and PS2R annealed to upon the initial PCR screening.

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
543
Therefore, taken together, these results demonstrated
that the biochemical pathway to matairesinol (5) in L.
the homologous region in the F. intermedia dirigent
(Fig. 6). The locations of the amino acid residues corre-
sponding to the forward and reverse degenerate primers
used to obtain this fragment are indicated in Fig. 7 by
bold, italicized, underlined characters. This fragment
was excised from the plasmid, labeled with 32P-dCTP,
¯
more, in the latter species, the matairesinol (5) so
avum was equivalent to that in Forsythia. Further-
0
formed is next hydroxylated to 7 -hydroxymatairesinol
(16), which is then converted into 5-methoxypodo-
5
phyllotoxin (4). The remaining conversions to this
important lignan thus involve aryltetrahydronaphtha-
lene ring formation, additional hydroxylation and O-
methylation steps, as well as formation of the methylene-
dioxy bridge. The latter can be assumed to result from a
NADPH-dependent cytochrome P-450 catalyzed ring
closure as demonstrated previously for the sesame (Sesa-
mum indicum) lignans (Jiao et al., 1998). Investigation of
the actual enzymology involved in these latter steps to
and used as a probe for screening approximately 3Â10 pfu
of the cDNA library, at a hybridization temperature of
ꢀ
55 C. After two rounds of screening, six positive clones
were excised into pBluescript phagemids using ExAssist
helper phage and transformed into SOLR cells (Strata-
gene). Of these six isolates, ®ve translated into an iden-
tical amino acid sequence (Fig. 8, PSD-Pp1) and the sixth
was similar except for two amino acid residues (Fig. 8,
PSD-Pp2). The 579 bp complete gene sequence (Fig. 7)
encodes a protein of 193 amino acid residues, including
a 40 amino acid secretory signal peptide; the calculated
mass of the secreted protein is ꢁ17 kDa. This amino
acid sequence has 68.4% similarity and 60.4% identity
to the F. intermedia dirigent (Fig. 8, PSD-Fi1) (psd-Fi1,
GenBank accession # AF210061) involved in a stereo-
selective phenoxy free-radical coupling reaction initiat-
ing the lignan biosynthesis pathway (Davin et al., 1997;
Gang et al., 1999a). Comparison of the peptides
excluding the signal peptide sequence shows an even
greater similarity between the two proteins (Fig. 8).
Three potential N-glycosylation sites are present (Fig. 7),
similar in number to the four sites in the Forsythia
dirigent gene (psd-Fi1).
5-methoxypodophyllotoxin (4) will be the subject of
additional studies.
Lastly, with the general pathway to podophyllotoxin
1) and 5-methoxypodophyllotoxin (4) so de®ned, it was
(
next instructive to obtain the gene encoding the dirigent
forming (+)-pinoresinol (6a) from P. peltatum, since
this represents the entry point into the overall pathway
and thus a promising candidate for genetic manipulation.
A rhizome cDNA library from P. peltatum was ®rst
constructed and screened for the corresponding dirigent
gene(s) as follows. Thus, a band of approximately 300
bp size was obtained by screening the P. peltatum rhizome-
derived cDNA library, using a polymerase chain reaction
(
PCR), with consensus primers designed from dirigent
homologs from 12 plant species (Davin and Lewis,
000). This 300 bp fragment was next cloned into a
The corresponding dirigent was next obtained in an
insect cell expression system capable of undergoing
posttranslational modi®cation. Primers were constructed to
produce a PCR fragment containing the clone for insertion
into a DES TOPO vector (Invitrogen) for transfection into
S2 Drosophila melanogaster cells. Selection of the trans-
fected insect cells in hygromycin medium yielded a
2
TOPO TA vector (Invitrogen), with plasmid prepara-
tions being procured from clones having the correct size
insert, and the plasmid DNA then sequenced. The
amino acid sequence corresponding to the 300 bp frag-
ment was found to be 72% identical and 80% similar to
Fig. 8. BOXSHADE comparison between entire P. peltatum dirigent homologues (PSD-Pp1 and PSD-Pp2) and F. intermedia dirigent (PSD-Fi1),
showing 68.4% similarity and 60.4% identity. Numbers to left of sequences indicate the amino acid residue locations. Black shading indicates that
the residue is identical to the column consensus; gray shading indicates that the residue is not identical but at least similar to the column consensus;
and no shading indicates that the residue is neither identical nor similar to the consensus (BOXSHADE version 3.2, written by Kay Hofmann and
Michael D. Baron). The presumed leader peptides are indicated by italicized characters.

5
44
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
population of cells that expressed the homologous diri-
gent: the gene produced the expected ꢁ26 kDa glycosy-
lated dirigent, which cross-reacted with F. intermedia
dirigent polyclonal antibodies (data not shown).
Promega's WizardTM Plus SV Minipreps DNA Pur-
i®cation System. DNA sequences were determined using
an Applied Biosystems model 373A automated sequen-
cer.
4
.2. Chemicals and enzymes
3. Concluding remarks
0
0
DEAE cellulose, 2 ,5 -ADP Sepharose and almond b-
In this study, it was demonstrated that the bio-
glucosidase were purchased from Sigma. FisherBiotech
brand of Thermostable Taq DNA polymerase was
obtained from Fisher Scienti®c and restriction endo-
nuclease enzymes were purchased from Promega.
TOPO TA, pTrcHis2 TOPO TA, and DES TOPO TA
cloning kits, TOP10 competent E. coli cells, and Droso-
phila melanogaster S2 cells were all obtained from
chemical pathway leading to formation of (� )-matai-
resinol (5b) in L. ¯avum parallels that occurring in
F. intermedia. Moreover, following its formation, this
0
intermediate is hydroxylated at the 7 -position to
0
a€ord 7 -hydroxymatairesinol (16), which is sub-
sequently converted into 5-methoxypodophyllotoxin
32
(
4). Thus, the biosynthetic pathway to this important
lignan involves dirigent mediated coupling to a€ord
+)-pinoresinol (6a), and subsequent metabolism to
give podophyllotoxin (1) and its derivatives. Addi-
tionally, heterologously expressed P. peltatum
dirigent cross-reacted with F. intermedia dirigent
antibodies, and thus is expected to also control coni-
feryl alcohol (7) derived coupling to give (+)-pin-
oresinol (6a), the initial step in the biosynthesis of
podophyllotoxin (1).
Invitrogen. Radiolabeled nucleotides [a- P]-dCTP and
32
[a- P]-dATP were purchased from DuPont NEN. Oli-
gonucleotide primers used for polymerase chain reac-
tion (PCR) experiments and for sequencing were
synthesized by Life Technologies, Inc. Puri®cation of
PCR fragments was accomplished using low melting
point (L.M.P.) agarose from Life Technologies, Inc.,
followed by enzymatic digestion with Promega's Agar-
ACE1 according to manufacturer's instructions.
(
4
.3. Plant material
4. Experimental
L. ¯avum plants, obtained from Hillview Gardens Pro-
ducts, Kennewick, WA, were maintained in Washington
State University greenhouse facilities. P. peltatum plants,
propagated from rhizomes harvested in Virginia, were
cultivated in the same greenhouse facilities.
4
.1. General procedures
1H NMR spectra were obtained on a Bruker
AMX300 spectrometer and chemical shifts are given in
ppm relative to TMS. Electron impact (EI) mass
ꢀ
4.4. Chemical syntheses
spectra were recorded on a Waters Integrity HPLC/MS
system at an ionization voltage of 70 eV, whereas high
resolution mass spectra were obtained on a VG 7070
EHF mass spectrometer. UV spectra and RNA/DNA
quanti®cation employed a Lambda 6 UV/VIS spectro-
photometer (Perkin-Elmer). Silica gel thin layer and
column chromatographic procedures were performed
using Kieselgel 60 F254 and silica gel 60 (EM Science,
0
4.4.1. 4,4 -Dibenzyl-7-oxo-matairesinol (14)
To a soln of thioacetal (11) (Katayama et al., 1993)
ꢀ
(1.0 g, 2.25 mmol) in dry THF under Ar at � 78 C, was
added dropwise n-BuLi (1.6 M in hexanes, 1.4 ml, 2.24
mmol) over 30 min. The resulting suspension was
allowed to stir for 1 h, following which a soln of 2(5H)-
furanone (0.16 ml, 2.26 mmol) in THF (10 ml) was
added dropwise. After stirring for another 60 min,
compound (12) (Brown and Daugan, 1987) (700 mg,
2.27 mol) in dry THF (10 ml) was subsequently added.
After stirring for 1 h, the suspension was allowed to
warm to room temp. with EtOAc (100 ml) slowly
230±400 mesh), respectively, whereas high performance
liquid chromatography utilized a Waters HPLC system
as described (Davin et al., 1992) using either reversed-
phase (Waters, NovaPak C18, 150Â3.9 mm inner dia-
meter), or chiral (Chiral Technologies Inc., Chiralcel
OD or Chiralcel OC, 240Â4.6 mm inner diameter)
columns, with detection at 280 nm. Radioactive samples
were analyzed in ScintiVerse (Fischer Scienti®c) and
measured using a liquid scintillation counter (Packard,
Tricarb 2000 CA).
An Amplitron II Thermolyne thermocycler (Barn-
stead/Thermolyne Corporation) was used for PCR
ampli®cation. Plasmid DNA for probes and for
sequencing was puri®ed from E. coli host cells using
added. The suspension was washed with H O (50 ml)
2
and satd brine (50 ml), then dried (Na SO ), ®ltered and
2
4
evaporated to dryness in vacuo to give crude (13) (1.8
g). The resulting residue was dissolved in 15% aq THF
solution (60 ml), whereupon HgO (2 g) was added
followed by dropwise addition of BF Et O (1 ml). The
reaction mixture was allowed to stir for 20 min, following
which CHCl3 (120 ml) was added, with the whole
washed twice with H O (60 ml). The CHCl solubles
3
2
2
3

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
545
+
were dried (Na SO ) and evaporated to dryness with the
2
125 (17). HRMS m/z: found 374.1353 [M] , calculated
1
4
resulting residue subjected to silica gel column chromato-
graphy (30Â2 cm) eluted with EtOAc±hexanes (1:2) to
give compound (14) (420 mg, 34% yield). UV lmax
MeOH nm: 314, 280, 230 and 209.; MS m/z (rel. int.):
for C H O : 374.1365; H NMR (CDCl ): ꢀ 2.55±2.65
(1H, m, H-8), 2.8±3.03 (3H, m, H-7, H-8 b), 3.77 (3H, s,
OMe), 3.81 (3H, s, OMe), 3.86±4.02 (2H, m, H-9 ),
4.32±4.64 (1H, m, H-7 ), 5.71 (1H, s, OH-Ar), 5.84 (1H,
s, OH-Ar), 6.54±6.85 (6H, m, Ar-H). 13C NMR
(CD OD): ꢀ 30.65, 35.99, 44.23, 46.46, 56.11, 56.14,
70.60, 74.74, 110.38, 113.80, 115.80, 119.46, 123.10,
130.37, 135.19, 146.10, 146.80, 148.68, 148.85 and 182.22.
20
22
7
3
0
0
0
5
1
52 [M]+ (51), 284 (28), 241 (100), 225 (31), 194 (25),
81 (48), 167 (24), 151 (68), 136 (46), 121 (25), 107 (30).
+
3
HRMS m/z: found 552.2123 [M] , calculated for
1
C H O : 552.2148; H NMR (CDCl ): ꢀ 2.98±3.25
3
4
32
7
3
(
OMe), 3.90 (3H, s, OMe), 4.04±4.16 (2H, m, H-9 ),
2H, m, H-7), 3.49±3.58 (1H, m, H-8), 3.71 (3H, s,
0
0
3
4.4.4. [7 - H]-7-Hydroxymatairesinol (16)
To (14) (50 mg, 0.09 mmol) in MeOH (5 ml), was
0
4.30±4.41 (1H, m, H-8 ), 5.04 (2H, s, CH ), 5.22 (2H, s,
CH ), 6.51±7.44 (16H, m, Ar-H). CNMR (CD COCD ):
2
13
3
� 1
slowly added NaB H (925 MBq, 7.6 GBq mmol ).
4
2
3
3
ꢀ
35.13, 45.48, 47.04, 55.30, 55.55, 68.56, 70.56, 70.80,
10.93, 112.20, 113.46, 114.06, 121.71, 123.20, 127.86,
28.17, 128.50, 128.66, 129.46, 131.07, 136.96. 149.78,
49.92, 153.30 and 177.02.
The resulting reaction mixture was stirred for 20 min, to
which was added unlabeled NaBH (5 mg, 0.13 mmol),
with the whole stirred for an additional 20 min. The
reaction mixture was next acidi®ed to pH 4.0 with 2 N
HCl. To this was added EtOAc (25 ml) with the whole
1
1
1
4
0
4
.4.2. 7 -Hydroxydibenzyl matairesinol (15)
then washed with H O (10 ml). The organic solubles
2
To compound (14) (200 mg, 0.36 mmol) in MeOH (20
ml) was slowly added NaBH (30 mg, 0.79 mmol). The
were dried (Na SO ), and evaporated to dryness in
2
4
vacuo. Following reconstitution of the residue in a
minimum amount of EtOAc, the resulting solution was
applied to a silica gel column (10Â2 cm), eluted with
4
resulting reaction mixture was stirred for 20 min and
acidi®ed to pH 4.0 with 2 N HCl. To this was added
EtOAc (100 ml), with the whole then washed with H O
3
EtOAc±hexanes (1:1 and 2:1) to a€ord [7- H]-(15) (48
2
(
40 ml). The organic solubles were next dried (Na SO ),
2
mg, 120 MBq, 2.5 MBq/mg, 95% yield). To a solution
3
4
and evaporated to dryness in vacuo. Following recon-
stitution of the residue in a minimum amount of EtOAc,
the resulting solution was applied to a silica gel column
of [7- H]-(15) (48 mg, 120 MBq) in EtOAc (10 ml), was
added 10% Pd/C (20 mg). The resulting suspension was
then stirred for 1 h at room temp. under an H atmo-
2
(
20Â2 cm), eluted with EtOAc±hexanes (1:1 and 2:1) to
sphere, following which the reaction mixture was
applied to a short silica gel column (4Â2 cm), eluted
with EtOAc. The EtOAc solubles were combined and
evaporated to dryness in vacuo. The resulting residue
a€ord (15) (190 mg, 95% yield). UV l MeOH
max
nm:280, 230 and 209; MS m/z (rel. int.): 554 [M]+ (5),
4
1
5
65 (4), 243 (42), 227 (50), 225 (41), 181 (40), 153 (74),
37 (100), 131 (61), 105 (39). HRMS m/z: found
was reconstituted in a minimum amount of Me CO,
2
+
1
54.2282 [M] , calculated for C H O : 554.2304; H
34
then applied to a silica gel column (10Â2 cm), and
34
7
0
3
0
NMR (CDCl ): ꢀ 2.5±2.66 (1H, m, H-8), 2.84±3.09 (3H,
eluted with EtOAc±hexanes (2:1) to yield [7 - H]-7 -
hydroxymatairesinol (16) (14.6 mg, 54 MBq, 3.7 MBq
3
0
m, H-7, H-8 ), 3.82 (3H, s, OMe), 3.84 (3H, s, OMe),
0
0
� 1
3
.84±3.96 (2H, m, H-9 ), 4.31±4.64 (1H, m, H-7 ), 5.11
mg , yield 45%).
(
1
2H, s, CH ), 5.15 (2H, s, CH ), 6.52±7.46 (16H, m, Ar-H).
2
2
3
C NMR (CD OD): ꢀ 29.57, 36.33, 43.84, 46.61, 56.25,
4.4.5. Syntheses of (Æ)-pinoresinols (6a/6b), (Æ)-secoiso-
3
0
3
7
1
1
1.14, 72.01, 72.10, 74.52, 110.68, 114.20, 114.85, 115.04,
18.88, 122.60, 128.70, 128.72, 129.32, 132.28, 136.91,
38.64, 138.70, 148.07, 148.46, 150.65, 150.70 and 181.05.
lariciresinols (9a/9b), (Æ)-[9,9 - H ) secoisolariciresinols
2
(9a/9b) and (Æ)-matairesinols (5a/5b)
These were prepared as previously reported (Ume-
zawa et al., 1991; Chu et al., 1993; Gang et al., 1999b).
0
4
.4.3. 7 -Hydroxymatairesinol (16)
To a solution of (15) (190 mg, 0.34 mmol) in EtOAc
40 ml), was added 10% Pd/C (100 mg). The resulting
4.5. Crude soluble protein preparation from L. ¯avum
(
suspension was stirred for 1 h at room temp. under an
Whole plant material (300 g) was frozen (liquid N2),
and pulverized in a Waring blender for 20 sec. The
resulting powder was homogenized in Tris±HCl bu€er
(50 mM, pH 7.5) containing 5 mM dithiothreitol (bu€er
A, 500 ml). The corresponding homogenate was next
®ltered through four layers of cheesecloth into a beaker
containing 10% (wt./vol.) polyvinylpolypyrrolidone.
The ®ltrate was centrifuged (12,000Âg, 15 min), with the
resulting supernatant employed for both enzyme assays
and protein puri®cation, respectively.
H atmosphere, following which the whole was applied
2
to a short silica gel column (4Â2 cm) eluted with EtOAc
to remove the Pd/C catalyst. The EtOAc solubles were
combined and evaporated to dryness, with the resulting
residue reconstituted in a minimum amount of Me CO,
2
and applied to a silica gel column (20Â2 cm), eluted
with EtOAc±hexanes (2:1) to give (16) (96 mg, 75%
yield). UV lmax MeOH nm: 284, 228 and 202; MS m/z
+
(
rel. int.): 374 [M] (23), 177 (20), 153 (100), 137 (54),

5
46
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
4.6. Partial puri®cation of L. ¯avum secoisolariciresinol
dehydrogenase
30:70 between 0 and 35 min; then to 5:95 in 5 min with
this being held at this composition for an additional 5 min,
at a ¯ow rate of 1 ml min 1. Fractions corresponding to
matairesinol (5) were individually collected, with aliquots
removed for liquid scintillation counting, and the
remainder freeze-dried for chiral HPLC analysis (Chiralcel
OD column) as described below.
�
The above crude enzyme preparation (150 mg protein
in bu€er A) was applied to a DEAE cellulose column
(
20Â2.6 cm) equilibrated in bu€er A. After washing the
column with bu€er A (25 ml), secoisolariciresinol dehy-
drogenase was eluted with a linear NaCl gradient (0±2
0
3
The optical activity of enzymatically formed [9 - H]-
matairesinol (5) was determined by chiral HPLC analysis
M in 500 ml) in bu€er A at a ¯ow rate of 2.5 ml min 1.
Active fractions were combined, concentrated via Ami-
con ultra®ltration (PM10 membranes) to ca 20 ml and
dialyzed against Tris±HCl bu€er (25 mM, pH 7.5 con-
taining 5 mM DTT, bu€er B). The dialyzed enzyme
solution (13 mg protein in bu€er B) was next applied to
�
0
3
after its chemical conversion to [9 - H]-secoisolariciresinol
0
3
(9). The freeze-dried enzymatically formed [9 - H]-
matairesinol (5) (0.5 kBq, 0.18 MBq mg 1) puri®ed
�
above was mixed with unlabeled matairesinol (5) (0.5
mg) in dry THF (2 ml), and reduced with LiAlH (1 mg)
4
0
0
a 2 ,5 -ADP-Sepharose column (10Â1 cm) equilibrated
in bu€er B. After washing with bu€er B, the column
was next eluted with 10 mM NAD in bu€er B. The
active fractions were combined, concentrated to 5 ml,
and assayed for matairesinol (5) formation.
for 1 h. The reaction was quenched with HCl (2 N, 20 ml),
and the resulting mixture was directly applied to silica gel
TLC. The TLC was performed in EtOAc±hexanes±MeOH
(10:10:1) with the secoisolariciresinol (9) band (R 0.25)
f
eluted with EtOAc (3 ml). The EtOAc solubles were
0
3
evaporated in vacuo to give [9 - H]-secoisolariciresinol
(9) (0.25 mg, 0.25 kBq) which was then subjected to
chiral HPLC analysis (Chiralcel OD column) as pre-
viously described (Dinkova-Kostova et al., 1996).
4
4
.7. Protein and enzyme assays
.7.1. Pinoresinol/lariciresinol reductase
Pinoresinol/lariciresinol reductase activity was assayed
2
as previously reported (Dinkova-Kostova et al., 1996).
In brief, each assay consisted of (Æ)-pinoresinols (6a/6b)
4.7.2.2. Enzymatic formation of [Ar- H]-matairesinol
2
(5). (Æ)-[Ar- H]-Secoisolariciresinols (9a/9b) (5.5 mmol),
(
5 mM in MeOH, 20 ml), the enzyme extract (100 ml),
synthesized as previously described (Umezawa et al.,
1991), were incubated with the partially puri®ed seco-
isolariciresinol dehydrogenase (ca 2 mg protein), Tris±HCl
bu€er (50 mM, pH 8.8, 10 ml) and NAD (40 mmol).
3
� 1
[
4R- H]-NADPH (10 mM, 6.79 kBq mmol , 20 ml) and
Tris±HCl bu€er (20 mM, pH 8.0, 110 ml). The enzy-
matically formed lariciresinol (8) and secoisolariciresinol
ꢀ
(9) were separated by reversed phase HPLC, indivi-
After 2 h incubation at 30 C with shaking, the mixture
was extracted twice with EtOAc (2Â10 ml). The EtOAc
dually collected, freeze dried and further analyzed using
chiral HPLC (Chiralcel OC and OD columns).
solubles were combined, evaporated to dryness in
vacuo, reconstituted in MeOH±3% HOAc in H O (1:1,
2
4
4
(
.7.2. Secoisolariciresinol dehydrogenase
0
600 ml) and an aliquot subjected to HPLC analysis. The
enzymatically formed matairesinol (5) (0.8 mg) was col-
lected, freeze-dried, and further submitted to mass
spectral analysis. MS m/z (rel. int.) 363 (0.6) 362 (2.9),
361 (9.2), 360 (M+ +2) (15.1), 359 (8.7), 358 (1.7), 139
(51), 138 (benzylic group with one deuterium) (100), 137
(44), 123 (22.3), 122 (9.2).
3
.7.2.1. Enzymatic formation of [9 - H]-matairesinol
5). Secoisolariciresinol dehydrogenase activity was
0
3
assayed by monitoring the formation of [9 - H]-matai-
0
3
resinol (5). Each assay consisted of (Æ)-[9,9 - H ]secoiso-
2
3
lariciresinols (9) (1.4 nmol in EtOH, 5 ml, 3.4Â10 Bq,
�
1
2.42 GBq mmol ), NAD (4 mM in 0.1 M KPi bu€er,
pH 7.5, 5 ml) and bu€er (50 mM Tris±HCl, pH 8.8, 440
ml). The enzymatic reaction was initiated by addition of
4.8. Administration of matairesinol (5) to L. ¯avum roots
the enzyme preparation (50 ml). After 2 h incubation at
ꢀ
3
0 C with shaking, the mixture was extracted with EtOAc
Root pieces (200 g fresh weight, 2 cm sections) from
5-month old L. ¯avum plants were incubated with a soln
of matairesinol (5) [0.28 mmol, 100 mg in EtOH±H O
2
(
(
(
500 ml) containing unlabeled (Æ)-matairesinols (5a/5b)
3 mg) as radiochemical carriers. After centrifugation
13,800Âg, 5 min), the EtOAc solubles were removed,
(1:10, 5.5 ml)] for 6 h. The tissues were then extracted
for 2 h with MeOH (2Â100 ml) under conditions of
re¯ux. After cooling to room temp., the MeOH solubles
were combined and evaporated to dryness, with the
residue incubated with b-glucosidase (20 mg, 2.85 units/
and the extraction procedure was repeated with EtOAc
(500 ml) but without addition of unlabeled radio-
chemical carriers. For each assay, the EtOAc solubles
were combined, evaporated to dryness in vacuo, recon-
stituted in MeOH±3% HOAc in H O (1:1, 200 ml), with
an aliquot (20 ml) subjected to reversed-phase HPLC
analysis with the following elution conditions: linear
gradient CH CN±3% HOAc in H O from 10:90 to
�
1
mg ) in NaAc±HOAc bu€er (100 mM, pH 5.5, 30 ml)
for 2 h. The resulting mixture was extracted with EtOAc
(2Â100 ml), with the organic solubles combined, dried
(Na SO ) and evaporated to dryness. The residue so
2
3
2
2
4

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
547
obtained was reconstituted in a minimum amount of
species (Davin and Lewis, 2000). The primers used for PCR
0
Me CO, and applied to a silica gel column (30Â2 cm),
were: 5 -KGTGTTYGAYGAYCCYATTACYBTWGA-
0
2
0
eluted with EtOAc±hexanes (1:2, 1:1 and 2:1). Fractions
eluting after those containing matairesinol (5) were
combined, evaporated to dryness and further analyzed
by HPLC with the elution conditions previously described
CAAC-3 (= forward primer PS6F) and 5 -GAAAT-
0
AAACATCTCCYTCAWATGMATCRGT-3 (=reverse
primer PS2R), where K=T/G, Y=C/T, B=C/T/G, W=
A/T, R=A/G, and M=A/C, S= C/G, and X=A/C/T/G.
0
6
in Section 4.7.2. The peak corresponding to 7 -hydroxy-
To screen for dirigent homologues, 3Â10 pfu of the
matairesinol (16) (elution volume: 14.3 ml), formed fol-
lowing administration of matairesinol (5) to L. ¯avum
roots, was collected, freeze-dried and subjected to mass
spectral analysis. MS m/z (rel. int.):374 [M]+ (40), 177
P. peltatum cDNA library were used as template com-
bined with 2.5 pmol each of PS6F forward and PS2R
reverse primers. The PCR experiments were performed
in a ®nal volume of 50 ml, containing a Fisher Company
bu€er supplied with the Taq DNA polymerase, which
provided a reaction mixture of 10 mM Tris±HCl, pH
8.3, 50 mM KCl, 0.1% vol/vol Triton X-100, 2.5 mM
(
16), 153 (100), 137 (62), 125 (18).
0
3
0
4
(
.9. Administration of [7 - H]-7 -hydroxymatairesinol
16) to L. ¯avum roots
MgCl , 0.2 mM of each dNTP, and 1.0 unit of Taq
2
DNA polymerase. Thermocycler settings for PCR
ꢀ
0
3
0
[
7 - H]-7 -Hydroxymatairesinol (16) (110 mg, 0.41
ampli®cations were 35 cycles of 1 min at 96 C, 2 min at
�
1
ꢀ
ꢀ
ꢀ
MBq, 3.7 MBq mg ) in EtOH±H O (1:12.5, 540 ml)
48 C, and 3 min at 70 C, with a 5 min 72 C ®nal
extension cycle. PCR products were resolved on 1% low
melting point agarose gels. Nucleotide sequence deter-
mination of the ampli®ed bands was accomplished by
2
was incubated for 6 h with 2 g of L. ¯avum roots
as described above. The resulting tissues were extracted
for 2 h with MeOH (2Â4 ml) under conditions of
re¯ux. The MeOH solubles were combined, evaporated
to dryness, and the residue was next incubated with
b-glucosidase (1 mg, 2.85 units mg 1) in NaAc±HOAc
bu€er (100 mM, pH 5.5, 10 ml) for 2 h. The mixture
was extracted with EtOAc (2Â20 ml), with the organic
solubles combined, dried (Na SO ) and evaporated
1
cloning the AgarACE -digested isolated agarose PCR
bands into a TOPO TA vector and transforming into
competent TOP10 E. coli cells according to Invitrogen's
instructions. Bacterial colonies were screened for inserts
using a PCR screening technique of colony selection
and inoculation directly into 25 ml of PCR mix con-
taining M13 Forward and M13 Reverse primers (Invi-
trogen). These primer sequences border the cloning site
of the TOPO TA vector. Clones containing the pre-
dicted insert size, determined by the locations of the
degenerate primer sequences to consensus regions of
known dirigents, were grown in a 4 ml Luria-Bertani (LB)
culture medium (Sambrook et al., 1994) containing
�
2
4
to dryness. The resulting residue was puri®ed by
HPLC as described above. The peak corresponding
to 5-methoxypodo-phyllotoxin (4) (elution volume:
3
0.5 ml) was collected, freeze-dried and submitted to
HPLC analysis with fractions collected every min for
radioactivity determination using liquid scintillation
counting.
carbenicillin (100 mg ml 1), processed into plasmid
DNA, and used for DNA sequence analysis. One clone
containing a sequence homologous to dirigent genes
previously isolated from other species was next selected
and used as a radiolabeled probe for cDNA library
screening. The probe was prepared by excising the
desired fragment out of the TOPO TA vector using
the Eco RI restriction enzyme sites, providing a DNA
fragment of 302 bp which included 11 bp of the TOPO
vector on each end of the fragment.
�
4
4
.10. Dirigent cloning and heterologous expression
.10.1. Podophyllum peltatum rhizome cDNA library
synthesis
Total RNA was extracted from fresh young rhizome
tissues using the LiCl precipitation procedure of Dong
1
and Dunstan (1996). A Promega PolyATtract mRNA
+
Isolation System was used to isolate poly(A) mRNA
from the total RNA. A P. peltatum rhizome cDNA
5
library, yielding a primary library titer of 6.5Â10 pfu,
was constructed using 5 mg of the mRNA with Strata-
gene's ZAP-cDNA synthesis kit, Uni-ZAP XR vector,
4.10.3. cDNA library screening
5
1
A total of 3Â10 pfu of the P. peltatum ampli®ed
and Gigapack1 III Gold Packaging Extract. The
library was ampli®ed once immediately, according to
cDNA library was plated with XL1-Blue MRF compe-
tent E. coli cells for primary screening according to
Stratagene's instructions. Phage plaques were lifted
onto Magna Nylon 137 mm membranes (Osmonics,
Inc.), air dried for at least 10 min, then ®xed and
denatured on the membranes using a one-step auto-
8
Stratagene's protocol, and gave a ®nal titer of 4.35Â10
�
1
pfu ml
.
4
.10.2. Dirigent DNA probe synthesis
Degenerate oligonucleotide primers used for dirigent
PCR screening were designed from the consensus
regions of 18 dirigent homologs obtained from 12 plant
ꢀ
claving procedure for 2 min at 100 C. The membranes
containing denatured plaques were next washed by gently
ꢀ
shaking for 30 min at 37 C in 6 X standard saline citrate

5
48
Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
(
SSC) (Sambrook et al., 1994) and then prehybridized
struct, along with the selection vector pCoHYGRO,
according to Invitrogen's instructions (Drosophila
Expression System, Version A). After 4 weeks selection
ꢀ
for 6 h with gentle shaking at 55 C in a covered crys-
tallization dish containing 300 ml of a hybridization
solution consisting of 6 X SSC, 0.5% SDS and 5 X
Denhardt's reagent. The probe was prepared by boiling
�
1
in 300 mg ml
HyQ SFX-Insect
hygromycin B (GIBCO BRL) in
TM
Serum-Free Insect Cell Culture
1
100 ng of the LMP-puri®ed DNA fragment, previously
digested from the TOPO TA vector plasmid prepara-
Medium (HyClone Company), a small-scale 20 ml pre-
paration of transfected cells was induced to express the
heterologous protein with CuSO at a ®nal concentra-
ꢀ
tion, for 10 min and then labeling for 20 min at 37 C
4
using Pharmacia's T7QuickPrime Kit and [a- P]-dCTP.
32
tion of 1 mM. Cells were harvested 36 h post-induction,
ꢀ
The radiolabeled probe was puri®ed from unin-
corporated nucleotides using a CENTRI SPINTM-20
.
centrifuged gently at 600Âg for 15 min at 4 C to pellet
the cells without damage, since this provides a super-
natant containing the secreted dirigent while minimizing
contamination by internal cytoplasmic proteins. The
supernatant was then centrifuged at 15,000Âg for 30
spin column (Princeton Separations, Inc) according to
the manufacturer's instructions. To the labeled probe
was added 0.5 mg sheared salmon sperm DNA (Sigma);
the sample was boiled again for 10 min, placed on ice
for 5 min, then added to 300 ml of fresh hybridization
solution which was preheated in a crystallization dish.
The membranes were transferred to this dish and
ꢀ
min at 4 C to remove any remaining debris. Ten micro-
grams of the ®nal supernatant were used for Western
analysis (Burlat et al., 2000) for detection of the secreted
dirigent protein.
ꢀ
allowed to hybridize at 55 C with gentle shaking. After
24 h, membranes were washed in 4 X SSC, 0.5% SDS
for 10±20 min at room temp. At this point, the mem-
branes were washed brie¯y in 6 X SSC at room temp.,
individually wrapped with plastic wrap, and exposed to
Acknowledgements
We gratefully acknowledge the National Science
Foundation (MCB9976684), the NSF REU Program,
and McIntire Stennis for ®nancial support.
ꢀ
Kodak X-OMAT AR ®lm for 24 h at � 80 C in a metal
cassette with intensifying screens. Positive plaques from
the ®rst screening were isolated through a second
screening using hybridization conditions similar to the
above.
References
4.10.4. In vivo excision and sequencing of P. peltatum
dirigent-containing phagemids
Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990.
Basic local alignment search tool. Journal of Molecular Biology 215,
4
03±410.
Puri®ed cDNA clones were rescued from phage using
the in vivo excision protocol outlined by Stratagene.
Both strands of the excised cDNAs were sequenced
completely.
Ayres, D.C., Loike, J.D., 1990. Chemistry and Pharmacology of Nat-
ural Products. Lignans: Chemical, Biological and Clinical Proper-
ties. Cambridge University Press, Cambridge, UK.
Beutner, K.R., von Krogh, G., 1990. Current status of podophyllo-
toxin for the treatment of genital warts. Seminars in Dermatology 9,
1
48±151.
4
.10.5. Sequence analysis
DNA and amino acid sequence analyses were per-
Broomhead, A.J., Rahman, M.M.A., Dewick, P.M., Jackson, D.E.,
Lucas, J.A., 1991. Matairesinol as precursor of Podophyllum lig-
nans. Phytochemistry 30, 1489±1492.
formed using the Unix-based GCG Wisconsin Package
Program Manual for the Wisconsin Package, 1997;
(
Brown, E., Daugan, A., 1987. The synthesis of (R)-(+)-b-vanillin-g-
butyrolactone and of chiral lignans therefrom. Heterocycles 26,
Rice, 1996). A homology search against the non-
redundant peptide database at the National Center for
Biotechnology Information was conducted using a
BLAST search program (Altschul et al., 1990).
1
169±1172.
Burlat, V., Kwon, M., Davin, L.B., Lewis, N.G., 2000. Monolignol par-
titioning in lignifying tissues: a role for dirigent proteins. (submitted).
Canel, C., Moraes, R.M., Dayan, F.E., Ferreira, D., 2000. Molecules
of interest: podophyllotoxin. Phytochemistry 54, 115±120.
4.10.6. Heterologous expression of dirigent in
Drosophila melanogaster cells
Chu, A., Dinkova, A., Davin, L.B., Bedgar, D.L., Lewis, N.G., 1993.
Stereospeci®city of (+)-pinoresinol and (+)-lariciresinol reductases
from Forsythia intermedia. Journal of Biological Chemistry 268,
Using gene-speci®c N- and C-terminal primers which
included the native start and stop codons, respectively, a
PCR product was generated using the pBluescript phage-
mid vector containing the dirigent clone as a template.
The PCR product, which included a putative native
secretory signal peptide, was then cloned into Invitrogen's
DES TOPO TA vector downstream of the inducible
metallothionine promoter. D. melanogaster S2 cells were
next transfected with the DES TOPO TA dirigent con-
2
7026±27033.
Davin, L.B., Bedgar, D.L., Katayama, T., Lewis, N.G., 1992. On the
stereoselective synthesis of (+)-pinoresinol in Forsythia suspensa
from its achiral precursor, coniferyl alcohol. Phytochemistry 31,
3869±3874.
Davin, L.B., Lewis, N.G., 2000. Dirigent proteins and dirigent sites
explain the mystery of speci®city of radical precursor coupling in
lignan and lignin biosynthesis. Plant Physiology 123, 453±461.
Davin, L.B., Wang, H.-B., Crowell, A.L., Bedgar, D.L., Martin,
D.M., Sarkanen, S., Lewis, N.G., 1997. Stereoselective bimolecular

Z.-Q. Xia et al. / Phytochemistry 55 (2000) 537±549
549
phenoxy radical coupling by an auxiliary (dirigent) protein without
an active center. Science 275, 362±366.
Program Manual for the Wisconsin Package. 1997. Genetics Com-
puter Group, Madison, WI, USA.
Dinkova-Kostova, A.T., Gang, D.R., Davin, L.B., Bedgar, D.L., Chu,
A., Lewis, N.G., 1996. (+)-Pinoresinol/(+)-lariciresinol reductase
from Forsythia intermedia: protein puri®cation, cDNA cloning,
heterologous expression and comparison to iso¯avone reductase.
Journal of Biological Chemistry 271, 29473±29482.
Rice, P., 1996. Program Manual for the EGCG Package. The Sanger
Centre. Hinxton Hall. Cambridge, UK.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1994. Molecular Cloning: A
Laboratory Manual, 3rd Edition. Cold Spring Harbor Laboratory,
Cold Spring Harbor, New York.
Dong, J.-Z., Dunstan, D.I., 1996. A reliable method for extraction of
RNA from various conifer tissues. Plant Cell Reports 15, 516±521.
Gang, D.R., Costa, M.A., Fujita, M., Dinkova-Kostova, A.T., Wang,
H.-B., Burlat, V., Martin, W., Sarkanen, S., Davin, L.B., Lewis,
N.G., 1999a. Regiochemical control of monolignol radical coupling:
a new paradigm for lignin and lignan biosynthesis. Chemistry and
Biology 6, 143±151.
Schacter, L., 1996. Etoposide phosphate: what, why, where and how?
Seminars in Oncology 23, 1±7.
Umezawa, T., Davin, L.B., Lewis, N.G., 1991. Formation of lignans,
(� )-secoisolariciresinol and (� )-matairesinol with Forsythia inter-
media cell-free extracts. Journal of Biological Chemistry 266, 10210±
10217.
van Uden, W., Pras, N., Malingre, T.M., 1990a. On the improvement
Â
Gang, D.R., Kasahara, H., Xia, Z.-Q., Vander Mijnsbrugge, K.,
Bauw, G., Boerjan, W., Van Montagu, M., Davin, L.B., Lewis,
N.G., 1999b. Evolution of plant defense mechanisms: relation-
ships of phenylcoumaran benzylic ether reductases to pinoresinol-
lariciresinol and iso¯avone reductases. Journal of Biological
Chemistry 274, 7516±7527.
of the podophyllotoxin production by phenylpropanoid precursor
feeding to cell cultures of Podophyllum hexandrum Royle. Plant Cell,
Tissue and Organ Culture 23, 217±224.
van Uden, W., Pras, N., Vossebeld, E.M., Mol, J.N.M., Malingre,
Â
T.M., 1990b. Production of 5-methoxypodophyllotoxin in cell sus-
pension culture of Linum ¯avum L. Plant Cell, Tissue and Organ
Culture 20, 81±87.
Jiao, Y., Davin, L.B., Lewis, N.G., 1998. Furanofuran lignan meta-
bolism as a function of seed maturation in Sesamum indicum:
methylenedioxy bridge formation. Phytochemistry 49, 387±394.
Katayama, T., Davin, L.B., Chu, A., Lewis, N.G., 1993. Novel
benzylic ether reductions in lignan biogenesis in Forsythia inter-
media. Phytochemistry 33, 581±591.
Williams, S.D., Birch, R., Einhorn, L.H., 1987. Treatment of dis-
seminated germ cell tumors with cisplatin, bleomycin and either
vinblastine or etoposide. The New England Journal of Medicine
316, 1433±1441.
Xia, Z.-Q., Costa, M.A., Davin, L.B., Lewis, N.G., 2000. Secoiso-
lariciresinol dehydrogenase puri®cation, cloning and functional
expression: implications for human health protection (submitted).
Young, R.C., 1992. Etoposide in the treatment of non-Hodgkin's
lymphomas. Seminars in Oncology 19 (Suppl. 13), 19±25.
Ziegler, F.E., Schwartz, J.A., 1978. Synthetic studies on lignan lac-
tones: aryl dithiane route to (Æ)-podorhizol and (Æ)-iso-
podophyllotoxone and approaches to the stegane skeleton. Journal
of Organic Chemistry 43, 985±991.
Krahmer, R.L., Hemingway, R.W., Hillis, W.E., 1970. The cellular
distribution of lignans in Tsuga heterophylla wood. Wood Science
and Technology 4, 122±139.
Lewis, N.G., Davin, L.B., 1999. Lignans: biosynthesis and function.
In: Barton, Sir D.H.R., Nakanishi, K., Meth-Cohn, O. (Eds.),
Comprehensive Natural Products Chemistry. Vol. 1. Elsevier, Lon-
don, pp. 639±712.
O'Dwyer, P.J., Leyland-Jones, B., Alonso, M.T., Marsoni, S., Wittes,
R.E., 1985. Etoposide (VP-16-213): current status of an active antic-
ancer drug. The New England Journal of Medicine 312, 692±700.