Studies on the late steps of (+) pisatin biosynthesis: Evidence for (-) enantiomeric intermediates

Article abstract of DOI:10.1016/j.phytochem.2005.12.027

Pisatin, a 6a-hydroxyl-pterocarpan phytoalexin from pea (Pisum sativum L.), is relatively unique among naturally occurring pterocarpans by virtue of the (+) stereochemistry of its 6a-11a C-C bond. However, pisatin synthesizing pea tissue has an isoflavone reductase, first identified in alfalfa, which acts on the (-) antipode. In order to establish the natural biosynthetic pathway to (+) pisatin, and to evaluate the possible involvement of intermediates with a (-) chirality in its biosynthesis, we administered chiral, tritium-labeled, isoflavanones and pterocarpans to pisatin-synthesizing pea cotyledons and compared the efficiency of their incorporation. Pea incorporated the isoflavanone, (-) sophorol, more efficiently than either its (+) antipode, or the pterocarpans (+) or (-) maackiain. (-) Sophorol was also metabolized by protein extracts from pisatin-synthesizing pea seedlings in a NADPH-dependent manner. Three products were produced. One was the isoflavene (7,2′-dihydroxy-4′,5′-methylenedioxyisoflav-3-ene), and another had properties consistent with the isoflavanol (7,2′-dihydroxy- 4′,5′-methylenedioxyisoflavanol), the expected product for an isoflavanone reductase. A cDNA encoding sophorol reductase was also isolated from a cDNA library made from pisatin-synthesizing pea. The cloned recombinant sophorol reductase preferred (-) sophorol(+) sophorol as a substrate and produced 7,2′-dihydroxy-4′,5′-methylenedioxyisoflavanol. Although no other intermediates in (+) pisatin biosynthesis were identified, the results lend additional support to the involvement of intermediates of (-) chirality in (+) pisatin synthesis.

Full text of DOI:10.1016/j.phytochem.2005.12.027

PHYTOCHEMISTRY
Phytochemistry 67 (2006) 675–683
www.elsevier.com/locate/phytochem
Studies on the late steps of (+) pisatin biosynthesis: Evidence for
(ꢀ) enantiomeric intermediates
Gregory L. DiCenzo , Hans D. VanEtten
1
*
Division of Plant Pathology and Microbiology, Plant Science Department, College of Agriculture, University of Arizona,
P.O. Box 210036, Tucson, AZ 85721-0036, USA
Received 11 August 2005; accepted 30 December 2005
Available online 28 February 2006
Abstract
Pisatin, a 6a-hydroxyl-pterocarpan phytoalexin from pea (Pisum sativum L.), is relatively unique among naturally occurring pterocar-
pans by virtue of the (+) stereochemistry of its 6a–11a C–C bond. However, pisatin synthesizing pea tissue has an isoflavone reductase,
first identified in alfalfa, which acts on the (ꢀ) antipode. In order to establish the natural biosynthetic pathway to (+) pisatin, and to
evaluate the possible involvement of intermediates with a (ꢀ) chirality in its biosynthesis, we administered chiral, tritium-labeled,
isoflavanones and pterocarpans to pisatin-synthesizing pea cotyledons and compared the efficiency of their incorporation. Pea incorpo-
rated the isoflavanone, (ꢀ) sophorol, more efficiently than either its (+) antipode, or the pterocarpans (+) or (ꢀ) maackiain. (ꢀ) Sop-
horol was also metabolized by protein extracts from pisatin-synthesizing pea seedlings in a NADPH-dependent manner. Three products
were produced. One was the isoflavene (7,20-dihydroxy-40,50-methylenedioxyisoflav-3-ene), and another had properties consistent with
the isoflavanol (7,20-dihydroxy-40,50-methylenedioxyisoflavanol), the expected product for an isoflavanone reductase. A cDNA encoding
sophorol reductase was also isolated from a cDNA library made from pisatin-synthesizing pea. The cloned recombinant sophorol reduc-
tase preferred (ꢀ) sophorol over (+) sophorol as a substrate and produced 7,20-dihydroxy-40,50-methylenedioxyisoflavanol. Although no
other intermediates in (+) pisatin biosynthesis were identified, the results lend additional support to the involvement of intermediates of
(ꢀ) chirality in (+) pisatin synthesis.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Pisum sativum L.; Leguminosae; Pea; Biosynthesis; Isoflavonoids; Pterocarpans
1. Introduction
(5) (Fig. 1A). Unlike pea, most legumes produce the (ꢀ)
stereoisomer of pterocarpans (Dewick, 1988) and have
the configuration at asymmetric carbons 6a and 11a
shown for (ꢀ) maackiain (3b) (Fig. 1A). The late steps
of pisatin (5) biosynthesis, particularly the steps that lead
to the (+) enantiomer and to hydroxylation at the 6a car-
bon, are not understood. Most models for pterocarpan
biosynthesis involve reduction of an isoflavone, catalyzed
by isoflavone reductase (IFR), as an intermediate step in
the biosynthesis of these compounds (Aoki et al., 2000;
Dixon, 1999).
The pterocarpan pisatin (5) (see Fig. 1) from garden
pea was the first chemically identified phytoalexin (Cru-
ickshank and Perrin, 1960) and is a product of the isofl-
avonoid pathway (Dixon, 1999). The basic pterocarpan
carbon skeleton contains rings A–D as shown for pisatin
(5) in Fig. 1A. In nature, pterocarpans exist in two stereo-
isomeric forms, one of which is exemplified by (+) pisatin
*
Corresponding author. Tel.: +1 520 621 1828; fax: +1 520 621 9290.
IFR was first detected in chickpea (Tiemann et al., 1987)
and cDNAs encoding IFR have been cloned from chick-
pea, alfalfa, pea and Lotus japonicus (Paiva et al., 1991,
E-mail address: vanetten@ag.arizona.edu (H.D. VanEtten).
Present address: Department of Veterinary Science and Microbiology,
1
University of Arizona, Tucson, AZ 85721, USA.
0031-9422/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2005.12.027

676
G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
The IFR that has been isolated and characterized from
alfalfa, a plant which makes only (ꢀ) pterocarpans, only
converts optically inactive isoflavones into (ꢀ) isoflava-
nones (Paiva et al., 1991), supporting the prediction of
Dewick and Ward that there is a specific reductase for
the synthesis of (ꢀ) pterocarpans. However, the IFR enzy-
matic activity detected in pea tissue that synthesizes (+)
pisatin (5) and the IFR encoded by an Ifr cDNA isolated
from this tissue, catalyze the conversion of the achiral iso-
flavone
7,20-dihydroxy-40,50-methylenedioxyisoflavone
(DMD) (1) to the (ꢀ) isoflavanone (ꢀ) ‘‘sophorol’’ (2b)
(Fig. 1A) (Paiva et al., 1994). In alfalfa, the final two
enzymes that form (ꢀ) medicarpin (3-hydroxy-9-methoxy
pterocarpan) from the (ꢀ) isoflavanone (vestitone:
7,20-dihydroxy-40-methoxyisoflavanone) are isoflavanone
reductase (vestitone reductase, VR), which acts on (ꢀ) ves-
titone to make (ꢀ) 7,20-dihydroxy-40-methoxyisoflavanol
(DMI), and the isoflavanol dehydratase that dehydrates
(ꢀ) DMI to make (ꢀ) medicarpin (Guo et al., 1994a,b)
(structures not shown). A cDNA encoding VR from alfalfa
has been cloned (Guo and Paiva, 1995).
Another uncertainty in the biosynthesis of (+) pisatin
(5) involves the origin of the oxygen in the hydroxyl group
at 6a. Radiolabeling studies in pea found that (+) maacki-
2
ain (3a) is readily incorporated into (+) pisatin (5). H
NMR analysis of pisatin (5), which had incorporated (+)
[11a-2H, 6-2H] maackiain (3a/b), revealed that both labeled
protons were retained, suggesting that hydroxylation at the
6a carbon probably involved an O2-dependent oxygenase
and did not desaturate either the 11a–6a or 6–6a carbon
bonds (Banks and Dewick, 1983a,b). However, subsequent
labeling experiments with 18O2 and H218O failed to support
this model for the origin of the 6a-hydroxyl group in pisa-
tin (5) and indicated that direct hydroxylation of maackiain
(3) is not involved in (+) pisatin (5) biosynthesis (Matthews
et al., 1987, 1989).
While the step(s) responsible for determining the stereo-
chemistry of (+) pisatin (5) have yet to be described, all of
the data indicate that the final step in pisatin (5) biosynthe-
sis is methylation of (+) 6a-hydroxymaackiain (4) at the
3-0 position (Fig. 1A). A methyl transferase, (+) 6a-
hydroxymaackiain 3-O-methyltransferase (HMM), that
catalyzes this reaction has been purified from pea (Preisig
et al., 1989). cDNA clones encoding HMM have been iso-
lated and used to make antisense (anti Hmm) constructs of
Hmm. When anti Hmm constructs are expressed in trans-
genic pea tissue (hairy roots), the tissue has a decreased
ability to produce pisatin (5) (Wu et al., 1997; Wu and
VanEtten, 2004).
The purpose of this present study was to evaluate fur-
ther the participation of the (ꢀ) and (+) isoflavonoid inter-
mediates in the synthesis of (+) pisatin (5). This was done
by measuring the relative incorporation of sophorol (2) and
maackiain (3) enantiomers into (+) pisatin (5) and by
determining if an enzyme (sophorol reductase) and a gene
analogous to (ꢀ) Vr from alfalfa are present in pea.
Fig. 1. (A) Proposed pathways for the biosynthesis of (+) pisatin (5) and
(ꢀ) maackiain (3b). The pathway proposed by Banks and Dewick (1982)
for (+) pisatin (5) is indicated by dashed arrows. Reactions for which the
enzymatic activity has been confirmed are indicated by solid arrows. The
occurrence of 7,20-dihydroxy-40,50-methylenedioxyisoflavanol (DMDI) (6)
as an intermediate in the (+) pisatin (5) pathway is indicated by a dotted
arrow. DMD = 7,20-dihydroxy-40,50-methylenedioxyisoflavone. (B) Struc-
ture of 7,20-dihydroxy-40,50-methylenedioxyisoflav-3-ene (7).
1994; Shimada et al., 2000; Tiemann et al., 1991). IFR is of
special significance, as the reduction of the 2,3 double bond
is the first reaction in the pterocarpanoid pathway that pro-
duces an asymmetric product. Dewick and Ward (1977)
proposed that this reaction could be the branchpoint lead-
ing to the two pterocarpan stereoisomers and that the IFR
leading to (ꢀ) pterocarpans, converts optically inactive iso-
flavones to (ꢀ) isoflavanones and the IFR leading to (+)
pterocarpans does the converse.

G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
677
2. Results and discussion
to wounded cotyledons, which had been treated with 5 mM
CuCl2. The (+) pisatin (5) synthesized during the incuba-
tion period was purified and the amount of tritium incor-
porated into it measured (Table 1). The time which
elapsed between addition of CuCl2 and the tritiated isofl-
avonoids and the length of precursor administration were
varied to give six different conditions. (ꢀ) Sophorol (2b)
was incorporated better than any of these other com-
pounds in all six conditions except for two individual treat-
ments in which there was no significant difference: (+)
sophorol (2a) in experiment A and (+) maackiain (3a) in
experiment F (Table 1). The incorporation of tritiated
DMD into (+) pisatin (5) was assayed one time and was
similar to that of (ꢀ) sophorol (2b) and both were better
incorporated than any of the other compounds tested
(Table 1).
2.1. In vivo labeling
Most previous studies of the incorporation of radiolabel
from possible precursors into pisatin (5) have used imma-
ture pea pods (e.g., Banks and Dewick, 1983a,b). We devel-
oped a method, which uses cotyledons from 4- to 5-day old
pea seedlings and allows quick production of pea tissue for
such studies. The cotyledons synthesize pisatin (5) upon
wounding and the amount produced is increased if CuCl2
is applied to the wounded tissue (Fig. 2). When tritiated
(ꢀ) sophorol (2b) was administered to such tissue, it was
rapidly incorporated into pisatin (5) and by 8 h little of
the added sophorol (2) was detectable (Fig. 3).
Tritiated (+) and (ꢀ) sophorol (2a) and (b) and (+) and
(ꢀ) maackiain (3a) and (b) were individually administered
Though Banks and Dewick (1983b) had demonstrated
that (+) maackiain (3a) is incorporated into (+) pisatin
(5) in vivo, the current study is the first to compare the
incorporation efficiencies of both isomers of sophorol (2)
and maackiain (3) in the same experiment. The superior
incorporation of (ꢀ) sophorol (2b) is consistent with a
model of pisatin (5) biosynthesis that depends on an iso-
flavone reductase, which produces the (ꢀ) isoflavanone
from DMD during the biosynthesis of pisatin (5) rather
than the (+) enantiomer and is consistent with the lack
of an isomerase that produces (+) sophorol (2a). As indi-
cated previously, the Ifr isolated from pea has been shown
to encode a protein that produces (ꢀ) sophorol (2b) from
DMD (1) (Paiva et al., 1994). Recently, Wu and VanEtten
(2004) have shown that pea hairy roots containing sense
and antisense constructs of pea Ifr have a reduced ability
to make pisatin (5), which is also consistent with a pisatin
(5) biosynthetic pathway that involves (ꢀ) sophorol (2b) as
an intermediate rather than (+) sophorol (2a) (Fig. 1A). In
addition, the observed better incorporation of (ꢀ) sophorol
(2b) compared to (+) maackiain (3a) would not be expected
if (+) maackiain (3a) were a normal intermediate in the
pathway at the position shown in Fig. 1A: (ꢀ) sophorol
(2b) is placed upstream of (+) maackiain (3a) in models
that propose the latter as an intermediate. As expected,
the incorporation of (ꢀ) maackiain (3b) into (+) pisatin
(5) was generally poor, consistent with its lack of involve-
ment in any of the proposed pathways for (+) pisatin (5)
biosynthesis.
Fig. 2. Production of pisatin (5) by pea cotyledons. Cotyledon sections
were prepared from 4-d old seedlings and the cut surfaces treated with
5 mM CuCl2 (+ CuCl2) or H2O (ꢀ CuCl2) as described in Section 4.
2.2. Metabolism of (ꢀ) sophorol (2b) by pea proteins
Since the administration experiments suggested that (ꢀ)
sophorol (2b) is an intermediate in (+) pisatin (5) biosyn-
thesis, protein extracts of CuCl2-treated pea seedlings were
assayed for enzymatic activity with (ꢀ) sophorol (2b) as a
substrate. When incubated with NADPH, three new com-
pounds, one of which accumulated to a significant amount,
were detected (Fig. 4). The major product (‘‘unkn 1’’) was
identified as 7,20-dihydroxy-40,50-methylenedioxyisoflav-3-
ene (7) (Fig. 1B); it had mobility (Rf = 0.72) on TLC plates,
Fig. 3. In vivo incorporation of (ꢀ) sophorol (2b) into pisatin (5). Twenty-
five nmol [3H] (ꢀ) sophorol (4.0 Ci/mol), in 10 ll of 2% DMSO and 5 mM
CuCl2 were administered to each of two 6-day old cotyledon sections.
Samples were collected and extracted into ethyl acetate as indicated in the
text. The ethyl acetate soluble compounds were separated by TLC and the
TLC plates subjected to autoradiography. The 3 lanes to the right contain
standards: S = [3H] sophorol (2), M = [3H] maackiain (3) and P = [14C]
pisatin (5).

678
G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
Table 1
Efficiency of incorporation of various isoflavonoids into (+) pisatin (5) in vivoA
% Incorporation by experiment
[3H] PrecursorB
(Time (h) of CuCl2 added, time of substrate administration, length of administration period)
A(0, 0, 4)
B(0, 0, 12)
C(0, 12, 4)
D(0, 0, 12)
E(0, 1, 16)
F(0, 3.5, 26)
DMD (1)
1.16 0.48ab
2.77 0.67a
1.23 0.22b
1.38 0.25ab
0.15 0.00b
(ꢀ) SOP (2b)
(+) SOP (2a)
(+) MAA (3a)
(ꢀ) MAA (3b)
1.58 0.44a
0.92 0.01a,b
0.32 0.01b
0.08 0.01b
2.63 0.25a
0.65 0.11b
1.43 0.19b
1.09 0.43b
1.94 0.26a
0.85 0.08b
1.18 0.08b
0.21 0.05c
1.37 0.18a
0.45 0.07c
0.59 0.10b,c
2.24 0.51a
0.81 0.62b
2.89 0.28a
Data are the means of the percent incorporation into (+) pisatin (5), ( ) the standard error. Means with different lower case letters are significantly
different at p = <0.05.
(A) Two cotyledon cores per sample; total, 4 per precursor; 7.5 nmol substrate administered at the time of pisatin (5) induction in 2% DMSO/5 mM
CuCl2; feeding period, 4 h. (B) Same as A except the feeding period, 12 h. (C) Same as A except the substrates were administered 12 h after induction for
4 h. (D) 1 cotyledon section per sample, total, 6 per precursor; 15 nmol substrate fed at the time of pisatin (5) induction in 2% DMSO/5 mM CuCl2;
feeding period, 12 h. (E) 1 cotyledon core per sample, total, 5 per precursor; 5 nmol substrate in 8 lL EtOH fed to tissue 1 h after pisatin (5) induction for
1 h; feeding period, 16 h. (F) 2 cotyledon sections per sample, total, 6 per precursor; 10 nmol substrate fed 3.5 h after pisatin (5) induction; feeding period,
26 h.
A
Pisatin was identified as (+) pisatin (5) by HPLC on a chiral column.
DMD, 7,20-dihydroxy40,50-methylenedioxyisoflavone (1), SOP, sophorol (2) and MAA, maackiain (3).
B
there was a slight reduction in the amount of (+) sophorol
(2a) recovered from the cell free extracts after the incuba-
tion period (data not shown).
2.3. Cloning of a (ꢀ) isoflavonone reductase from pea
The radiolabeling studies and the activity of the cell free
extracts of CuCl2-treated pea tissue, are consistent with (ꢀ)
sophorol (2b) as an intermediate in the biosynthesis of pisa-
tin (5). Therefore an attempt was made to isolate a pea
gene encoding an enzyme, which would be analogus to
Vr and reduce (ꢀ) sophorol (2b). The Vr cDNA from
alfalfa (Guo and Paiva, 1995) was used as a heterologous
probe to screen a cDNA library made from pea seedlings,
which had been infected with Nectria haematococca. Sev-
eral clones hybridized strongly to the Vr probe and the
insert of one of these (p11-1) was sequenced. The 1241-
Fig. 4. Thin layer chromatogram (photographed under UV) of products
produced when (ꢀ) sophorol (2b) is incubated with crude protein extracts
bp insert of p11-1 encodes a 978-bp open reading frame,
and includes putative translation start and stop sites. The
nucleotide sequence of the open reading frame is 83.5%
identical to the Vr gene of alfalfa. The deduced amino acid
sequence of the pea (ꢀ) isoflavanone reductase gene, which
is called sophorol reductase (Sor), is 86.5% identical and
92.0% similar to Vr (Fig. 5).
from pea seedlings which had been induced to synthesize pisatin (5). All
reactions contained 50 lM sophorol (2), 2 mM NADPH, pH 6.7 and were
carried out at 30 ꢁC as described in Section 4. Reactions were stopped at
the indicated times. Included as standards were: maackiain (3), pisatin (5),
HMK, 6a-hydroxymaakiain (4), and sophorol (2). The protein extracts
from pea contained a small amount of pisatin (5).
HPLC retention time (Rt), UV spectrum, and mass spec-
trum as for the authentic standard (see Section 4). The sec-
ond compound (‘‘unkn 2’’) occurred transiently and was
detected by TLC only in reaction products from early time
points (Fig. 4). This compound had the same fluorescence
color under UV light and TLC mobility as 7,20-dihy-
droxy-40,50-methylenedioxyisoflavanol (6) (Fig. 1A). This
is the expected product if pea has an (ꢀ) isoflavanone
reductase equivalent to VR. A third compound (‘‘unkn’’ 3)
(Fig. 4) was not definitively identified. The production of
all three compounds from (ꢀ) sophorol (2b) was dependent
on the presence of NADPH. The compounds were not pro-
duced when (+) sophorol (2a) was substrate, although
2.4. Northern hybridization analysis
If Sor is involved in pisatin (5) biosynthesis, the induc-
tion of this gene would be expected to correlate with its
production and the expression of other genes, such as Ifr
and Hmm, known to be involved in pisatin (5) biosynthesis.
Four-day old pea seedlings were treated with CuCl2 to
induce pisatin (5) biosynthesis and total RNA was isolated
at various times after the treatment. The Sor probe hybrid-
ized with a RNA of approximately 1.35 kb (Fig. 6). A small
amount of Sor-hybridizing RNA was present prior to
CuCl2 treatment (lane 1) but increased significantly at later

G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
679
Fig. 5. Comparison of the deduced amino acid sequences from the sophorol reductase from pea and the vestitone reductase from alfalfa. Identical amino
acids are shaded.
Fig. 6. Northern hybridization analysis of Sor expression in pea. Total
RNA from pea seedlings was resolved in a formaldehyde gel, transferred
to Hybond-N+ membrane, and hybridized to 32P-labeled insert from
p3B1, a cDNA encoding SOR. (A) Autoradiogram, (B) the formaldehyde
gel stained with ethidium bromide. Lane 1: no CuCl2 treatment, lanes 2–9:
3, 6, 9, 12, 18, 21, 24, and 30 h post CuCl2 treatment, respectively. Size
standards (kb) are indicated on the right side in A.
time points, a pattern of accumulation which coincides
with pisatin (5) production and the accumulation of Ifr
and Hmm transcripts in CuCl2 – elicited pea tissue (Paiva
et al., 1994; Wu et al., 1997).
2.5. Genomic analysis
Fig. 7. Southern hybridization analysis of pea genomic DNA. Genomic
DNA from pea seedlings was digested with the restriction enzymes EcoRI
The Sor cDNA contains one site each for EcoRI and
HindIII, and two DraI sites. There are no BamHI sites in
the cDNA. In Southern hybridization analysis of genomic
DNA digested by EcoRI, HindIII, EcoRI/HindIII, DraI
and BamHI, a single DNA fragment hybridized in the lane
containing the BamHI digested DNA (Fig. 7), consistent
with one genomic copy of Sor. A strongly hybridizing
ꢁ800-bp DraI fragment is consistent with the predicted size
of an internal DraI fragment from the cDNA (819 bp) and
the hybridizination pattern of the HindIII and EcoRI/Hin-
dIII digests is consistent with the presence of HindIII and
EcoRI sites detected in the cDNA. The reason why only
(lane 1), HindIII (lane 2), EcoRI and HindIII (lane 3), DraI (lane 4), and
BamHI (lane 5), transferred to nylon membrane, and hybridized with 32P-
labeled p3B1 insert, a cDNA encoding SOR.
one, rather than the two fragments expected, hybridized
to EcoRI-digested DNA is unknown, but the results are
consistent with a single copy of Sor in pea.
Primers specific to the 50 translational start and 30 termi-
nation sites as well as to two internal sequences of the
cDNA were used to amplify DNA specific to the Sor gene
from genomic DNA and cloned cDNA. The PCR products

680
G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
with crude protein extracts of pea seedlings indicate that
7,20-dihydroxy-40,50-methylenedioxyisoflav-3-ene (6) and
the unidentified compound (unkn 3, Fig. 4) produced from
(ꢀ) sophorol (2b) also are made when DMDI (6) is the sub-
strate (G. DiCenzo and H. VanEtten, unpublished results).
The production of 7,20-dihydroxy-40,50-methylenedioxyi-
soflav-3-ene (6) is intriguing. This isoflavene is an achiral
molecule which could be converted subsequently to the
(+) asymmetry by hydration across the C3–C4 double
bond (Fig. 1B). However, radiolabeling studies have shown
that 7,20-dihydroxy-40,50-methylenedioxyisoflav-3-ene (7) is
incorporated poorly into (+) pisatin (5) and unlikely to be
an intermediate (Banks and Dewick, 1983). In contrast,
radiolabeled isoflavene is incorporated readily into coume-
stans and appears to be an intermediate in the synthesis of
these compounds (Martin and Dewick, 1980). While fur-
ther studies are required to identify the final steps in the
biosynthesis of (+) pisatin (5), the current results support
the involvement of compounds with the (ꢀ) stereochemis-
try as intermediates in this pathway.
Fig. 8. Metabolism of (ꢀ) sophorol (2b) by SOR. The cDNA contained in
p3B1, which encode sophorol reductase, was expressed in E. coli and used
to metabolize (ꢀ) sophorol (2b) (50 lM) in the absence (A) or presence (B)
of 1 mM NADPH for 30 min at 30 ꢁC.
made were of the same size whether the cDNA or genomic
DNA served as template. It appears, therefore, that the Sor
structural gene is identical to the coding region of the
cDNA and that no introns are present.
2.6. Analysis of recombinant SOR activity
Two cloned PCR products of Sor (from p2B1 and p3B1)
were expressed in E. coli and the SOR activity in the crude
enzyme preparations was compared to that of the recombi-
nant VR from alfalfa. The protein preparations from both
clones catalyzed the reduction of (ꢀ) sophorol (2b) to 7,20-
dihydroxy-40,50-methylenedioxyisoflavanol (DMDI) (6) in
a NADPH dependent manner (only data from protein of
p3B1 is shown in Fig. 8), identical to the activity of recom-
binant VR from alfalfa (data not shown). A trace amount
of DMDI (6) was also produced by all three proteins on
(+) sophorol (2a), which may be due to contamination of
the (+) sophorol (2a) stock with (ꢀ) sophorol (2b) caused
by racemization (Guo et al., 1994a).
4. Experimental procedures
4.1. Plant materials
Seeds of P. sativum L. cultivar Alaska were obtained
from Royal Seed, St. Joseph, MO. Cotyledons were pre-
pared for use in the incorporation studies in the following
way: The seeds were immersed in EtOH/H2O/5.25%
sodium hypochlorite (75:15:10) for 1 h, drained, rinsed 5·
with water, and left in water to imbibe for 3–4 h. After
imbibition, the seeds were spread over autoclaved moist
vermiculite and germinated at room temperature for 4–6
d in the dark. Cotyledons usually were removed from the
seedlings when the shoots were beginning to expand and
elongate. A section of cotyledon approximately 2 mm thick
was prepared for CuCl2-induction and [3H] substrate
administration by removing and discarding an ꢁ2-mm
thick dome-shaped section from the convex portion of
the cotyledon. Experiments were performed with the ca.
2 mm thick remaining section.
3. Conclusions
The better incorporation of (ꢀ) sophorol (2b) than (+)
sophorol (2a) into (+) pisatin (5) and the identification of
a (ꢀ) sophorol reductase, which is induced during the pro-
duction of (+) pisatin (5), are consistent with the involve-
ment of (ꢀ) sophorol (2b) as an intermediate in the
synthesis of (+) pisatin (5). It would appear that the step(s)
responsible for the asymmetry at carbon-6 and carbon-11
in (+) pisatin (5) occurs after formation of DMDI (6)
(Fig. 1, A). The optical form of the DMDI (6) produced
from (ꢀ) sophorol (2b) was not identified, but the (ꢀ) iso-
mer 6 is more likely as it is a reductive product of (ꢀ) sop-
horol (2b). Thus, in pea the synthesis of (+) pisatin (5)
apparently involves an unexpected stereochemical reversal
in which the first stereochemically active intermediates
have a stereochemistry associated with the synthesis of
(ꢀ) pterocarpans and the later intermediate(s) and final
product have the (+) stereochemistry.
To produce elicited pea tissue for enzyme assays, 4–6-
day-old seedlings produced as above, were immersed in
1 mM CuCl2 for 1 h and incubated in the dark for 24 h.
The seedlings were rinsed well and the cotyledons removed.
The seedlings were frozen in N2 and stored at ꢀ80 ꢁC.
4.2. Radiolabelling studies
One to two sections of cotyledon were treated with
5 mM CuCl2, either prior to or at the time of [3H]-isoflavo-
noid administration. Five to 35 ll of the candidate
precursor in either 2% DMSO with 5 mM CuCl2 or in
EtOH–H2O (1:1) was administered to the cotyledon sec-
tions. The time of addition of the candidate precursor var-
ied by experiment and ranged from 0 to 31 h post
This study did not identify the steps that follow the pro-
duction of DMDI (6). However, preliminary experiments

G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
681
induction. The tissue sections were incubated with the sub-
strate in several ways. In some experiments, the cotyledon
slice (described above in Section 4.1) was kept on water-
dampened Whatmann #4 filter paper in 35 mm polystyrene
petri dishes. In others, the cotyledons were trimmed with a
cork borer to ꢁ5 mm diameter and placed in 1.5-ml micro-
fuge tubes or in 16-mm glass test tubes. The incorporation
of radiolabeled substrates was stopped by freezing the coty-
ledons at ꢀ20 ꢁC. To prepare the frozen tissue for analysis,
it was macerated with a glass rod in 1 ml of 1 M Tris, pH
7.5 and the slurry extracted 2· with EtOAc. The organic
phases from each treatment were combined and concen-
trated under N2 or in a Speed Vac Concentrator (Savant
Instruments Inc, Farmingdale, NY). Pisatin (5) was sepa-
rated from the other components in the extract by silica
gel TLC and reversed phase HPLC and analyzed for incor-
poration of the radiolabel.
7,20-Dihydroxy-40,50-methylenedioxyisoflavonol (DMDI)
(6) was made from (ꢀ) sophorol (2b) by the method of Guo
and Paiva (1995). Briefly, a culture of E. coli DH5a harbor-
ing the pVR1 cDNA in the expression vector pSE380 was
grown at 37 ꢁC to an OD of 0.6 (A600). The expression of
the recombinant VR was induced with 5 mM isopropyl-b-
D-thiogalactopyranoside and the culture grown for an
additional 2 h at 37ꢁC. A reaction mixture (500 ll total vol-
ume) containing 50 ll cell lysate with 50 lM (ꢀ) sophorol
(2b), 1 mM NADPH in 0.2 mM phosphate pH 6.0 was
incubated at 30 ꢁC for 15 min; a longer period of incuba-
tion did not increase the yield of DMDI (6). DMDI (6)
was separated from the substrate by reversed phase HPLC
(see below). 7,20-Dihydroxy-40,50-methylenedioxyisoflav-3-
ene (7) was made following the method of Banks and
Dewick (1983b). b-NADPH was from Sigma Chemical
Company, St. Louis, MO.
4.3. Chemicals
4.4. Analysis of isoflavanone metabolites by TLC and HPLC
(+) and (ꢀ) Maackiain (3a) and (3b), (+) and (ꢀ) pisatin
(5) and [14C] pisatin were from laboratory stocks prepared
as described previously (George et al., 1998). DMD (7,20-
dihydroxy-40,50-methylenedioxyisoflavone) (1) was a gift
from Yuegin Sun (Eli Lilly and Company, Indianapolis,
IN 46285). [3H] maackiain (3) and DMD (1) were made
by a method based on Banks and Dewick (1983b) as fol-
lows: 25 mg (+) or (ꢀ) maackiain (3a) or (3b) or DMD
(1) was dissolved in 300 ll dimethylformamide to which
TLC and autoradiography were carried out essentially
as described by Preisig et al. (1990). The extracts were
applied to silica gel TLC plates (EM Sciences Silica gGel
60 F254, aluminum backed) and developed in tolu-
ene:EtOAc, 60:40. The developed plates were photo-
graphed under both long and short wavelength UV light
to record the locations of the compounds. Autoradiogra-
phy was carried out by coating the developed plate with
a 10% solution of 2,5-diphenyloxazole (PPO) in Et2O.
The solution was applied by painting the plate with a
wad of silanized glass wool dipped in the liquid. After dry-
ing, the plate was wrapped in polyethylene film and
enclosed in an X-ray cassette with Kodak X-Omat film
(Kodak, Rochester, NY). The X-ray film was exposed at
ꢀ80 ꢁC for 2–3 weeks before processing. In some experi-
ments, the compounds detected under both long and short
wavelength UV light, were extracted from the silica gel
with EtOH. Their UV absorbance spectra was measured
3
was added 100 ll H2O (500 mCi, Amersham) and 9.8 ll
triethanolamine. The solution was heated to 80 ꢁC for
ꢁ48 h after which it was acidified with 70 ll of 1 M HCl.
The solution was extracted with CHCl3 (2 · 1 mL) and
the combined CHCl3 fractions were washed with H2O
(2 ml). The [3H] labeled products were then purified by
using the following solvent systems sequentially: (1) hex-
ane–EtOAc, 3:2; (2) hexane–EtOAc–MeOH, 60:40:1; and
(3) hexane–Me2CO 2:1. While not determined in our exper-
iments, Banks and Dewick (1983b) showed that this proto-
col yields maackiain (3) with the majority of the [3H] on
C4.
(+) and (ꢀ) Sophorol (2a) and (2b) were obtained by
biotransformation of (+) and (ꢀ) maackiain (3a) and
(3b), respectively, by Colletotrechum trifolii (VanEtten lab
isolate T-456). Glucose-asparagine (GA) medium (100 ml)
(Matthews et al., 1987) was inoculated with 0.3 g fresh
weight (f.wt.) mycelia and conidia from an overnight GA
culture, which had been started from cultures of C. trifolii
grown on V-8 (Miller, 1955) agar slants. Maackiain (3) was
added to a final concentration of 30 lM and the culture
incubated at 25 ꢁC on a shaker (175 rpm). After 5–6 h,
the culture was extracted 2· with equal volumes of EtOAc.
The organic layers were combined and concentrated in a
with
a Beckman (Fullerton, CA) DU-64 scanning
spectrophotometer.
Reversed phase HPLC (RP-HPLC) used a gradient sys-
tem (24–74% CH3CN in H2O). Normal phase HPLC was
performed isocratically using MeCl2–EtOAc–EtOH–
AcOH (100:4.5:0.5:0.5). Both RP-HPLC and normal phase
HPLC runs were monitored at 309 and 278 nm. The UV
absorbance of selected compounds was measured as above.
Chiral chromatography was performed isocratically on
a ChiralPak OT+ column (J.T. Baker, Phillipsburg, PA)
(2.5 · 250 mm) in MeOH at 0.5 ml/min at 5 ꢁC. Absor-
bance was monitored at 309 and 278 nm. In this system,
the Rt for (ꢀ) pisatin was ꢁ9.3 and 10.2 min for (+) pisatin
(5).
Buchi RE121 rotary evaporator (Buchi, Flawil Switzer-
4.5. Analysis of isoflavanone metabolites by MS
¨
¨
land). Sophorol (2) was separated from maackiain (3)
and 1a-hydroxymaackiain (another product of maackiain
metabolism) by silica gel TLC in toluene:EtOAc (60:40).
The mass spectra of HPLC-purified samples were deter-
mined on an HP 5988A using a direct insertion probe, a

682
G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
temperature range of 30–350 ꢁC (ramped 30ꢁ/min) and an
EI of 70 eV. Alternatively, direct inlet was used on a Jeol
HX-110A Sector mass spectrometer in either CI+ mode
with iso-butane, or in EI+ mode. Derivatization was per-
formed using bis (Trimethyl-Sylil) Triflouroacetamide
(BSTFA) (Sigma) essentially according to the manufac-
turer’s instructions: the sample was dissolved in 14 ll
pyridine, 7 ll BSTFA were added and the mixture heated
to 70 ꢁC for 15 min.
using (50-ACG CGG ATC CTT AGA GAT ATC CTT
TTT CCT-30). Each PCR reaction contained PCR reaction
buffer, 40 ll H2O, each dNTP [200 lM], MgCl2 [1.5 mM],
primers [1.0 lM], 20 ng p11-1 (template), and 1.25 U Taq
polymerase (GibcoBRL) in a total volume of 50 ll. The
program was as follows: 2 min denaturation at 95 ꢁC, fol-
lowed by 35 cycles of: denaturation (95 ꢁC, 1 min), anneal-
ing (60 ꢁC, 1 min), and extension (72 ꢁC, 2 min). The
combination of primers produced an ꢁ1-kb DNA frag-
ment, which was gel-purified, extracted with phenol–
CHCl3, and precipitated with 1/10 vol. of 3 M NaOAc + 2
vol. of 100% EtOH. Two such PCR products, identified as
pcr 2B and pcr 3B, were purified and each was double-
digested with NcoI and BamHI. The expression vector
pSE380 was similarly digested and, after gel-purification
of the PCR products and the vector using the Qiaex II sys-
tem, the PCR fragments were ligated into pSE380. Two
clones, p2B1 and p3B1, containing PCR products 2B and
3B, respectively, were selected for further analysis.
4.6. Chemical properties of enzymatic products
7,20-Dihydroxy-40,50-methylenedioxyisoflav-3-ene
(7)
had the broad UV absorbance maximum centered at
ꢁ338 nm, and a GC–MS spectrum with major ions at
m/z = 283 (22%) and m/z = 284 (100%). The trimethylsylil
(TMS)-derivative yielded the expected molecular ion at m/
z = 428 (284 + 2 TMS). All of these values also were
obtained with the authentic standard.
4.7. Isolation of the cDNA of Sor
4.8. Northern analysis
A cDNA library (7.5 · 104 recombinants) of poly (A+)
RNA from pea tissue infected with N. haematococca in
the vector Uni-ZAP XR (Stratagene) (Han et al., 2001)
was screened by in situ plaque-hybridization with the insert
of pVR1 (gift of Nancy Paiva, Nobel Foundation, Ard-
more OK). Hybridization was performed at 65 ꢁC. The
blots were washed at 65 ꢁC with 1· SSC, 0.1% SDS for
45 min, and with 0.1· SSC, 0.1% SDS for 30 min. Fourteen
strongly-hybridizing plaques were subjected to a second
round of screening. The cDNA inserts were rescued in
pBluescript SK(ꢀ) following the in vivo excision protocol
for kZAP (Stratagene). DNA from each clone was digested
with EcoRI and XhoI and subjected to Southern analysis
using the insert of pVR1 as a probe. Several clones had a
strongly hybridizing 0.75-kb EcoRI fragment and a 0.5-
kb EcoRI/XhoI fragment, but also contained additional
insert DNA, suggestive of concatenated cDNAs. The
0.75- and 0.50-kb fragments from p11-1 were gel-purified
using the Qiaex II system (Qiagen, Santa Clarita, CA)
and ligated into pBS digested with EcoRI and XhoI. The
0.75-kb and the 0.50-kb inserts in the resulting plasmids
and the 50 and 30 ends of the insert in p11-1were sequenced
by DNA Sequencing Service, Arizona Research Laborato-
ries, University of Arizona. The sequence of the region
spanning the junction between the 0.75- and 0.05-kb frag-
ments was obtained from p11-1. The sequence of the full-
length cDNA was assembled from the partial sequences
(Genbank Accession Number AF107404).
Pea seedlings were treated with CuCl2 for 1 h and subse-
quently incubated for various lengths of time. RNA was
isolated from CuCl2-induced seeding using TRIZOL
Reagent (GibcoBRL Life Technologies, Gaithersburg,
MD) following the manufacturer’s instructions. Electro-
phoresis of isolated RNA in 1% agarose/formaldehyde gels
was performed as described by Ausubel et al. (1993). The
RNA was transferred onto nylon membranes (Hybond
N+ Amersham Biosciences Co., Piscataway, NJ), and
3
hybridized with P-labeled Sor obtained from p3B1. The
hybridization and the washing of the membrane were as
described for the isolation of the Sor cDNA.
4.9. Southern hybridization
Total genomic DNA was purified from 30 g of pea seed-
lings as described Saghai-Maroof et al. (1984). Genomic
DNA (100 lg) was digested with restriction enzymes, sub-
jected to electrophoresis in an 0.8% agarose gel and trans-
ferred onto nylon membrane (Hybond-N+). The blotted
DNA was prehybridized for 1 h at 65 ꢁC and subsequently
hybridized with the 32P-labeled probe for an additional 12–
18 h at 65 ꢁC. The membranes were washed as described
for the isolation of the Sor cDNA.
4.10. PCR analysis
PCR amplification of genomic DNA was compared to
that of the cDNA by using several combinations of prim-
ers: the primers specific to the 50 and 30 ends of the deduced
coding region (described above) were used to amplify a
992-bp fragment of the cDNA and the corresponding
region from genomic DNA. The 50 primer also was com-
bined with an internal downstream primer 852 bp after
the start of the cDNA coding region. Likewise, the primer
In order to clone a full-length homologue of the pVR1
cDNA from p11-1, synthetic primers (GibcoBRL) that
were specific to the predicted 50 and 30 ends of the cDNA
and that introduced unique restriction sites were utilized.
A NcoI site was introduced at the 50 end of the cDNA using
(50-ACG CCA TGG CAG AGG GGA AAG GAA GGG
T-30). Similarly, a BamHI site was introduced at the 30 end

G.L. DiCenzo, H.D. VanEtten / Phytochemistry 67 (2006) 675–683
683
specific to the 30 end of the gene was combined with an
upstream primer starting 688 bp from the ATG at the start
of the cDNA coding region. Both the p11-1 cDNA and
genomic DNA were used as templates in order to detect
sequence differences such as might arise from an intron.
George, H.L., Hirschi, K., VanEtten, H.D., 1998. Biochemical properties
of the products of cytochrome P450 genes (PDA) encoding pisatin
demethylase activity in Nectria haematococca. Arch. Microbiol. 170,
147–154.
Guo, L., Paiva, N.L., 1995. Molecular cloning and expression of alfalfa
(Medicago sativa L.) vestitone reductase, the penultimate enzyme in
medicarpin biosynthesis. Arch. Biochem. Biophys. 320, 353–360.
Guo, L.N., Dixon, R.A., Paiva, N.L., 1994a. The pterocarpan synthase of
alfalfa – association and coinduction of vestitone reductase and 7,20-
4.11. Preparation of 32P-labeled probes
dihydroxy-40-methoxy-isoflavanol (DMI) dehydratase, the
enzymes in medicarpin biosynthesis. FEBS Lett. 356, 221–225.
Guo, L., Dixon, R.A., Paiva, N.L., 1994b. Conversion of vestitone to
medicarpin in alfalfa (Medicago sativa L.) is catalyzed by two
independent enzymes. J. Biol. Chem. 269, 22372–22378.
Han, Y., Liu, X., Benny, U., Kistler, H.C., VanEtten, H., 2001. Genes
determining pathogenicity to pea are clustered on a supernumerary
chromosome in the fungal plant pathogen Nectria haematococca..
Plant J. 25, 305–314.
Martin, M., Dewick, P.M., 1980. Biosynthesis of pterocarpan, isoflavan
and coumestan metabolites of Medicago sativa – the role of an isoflav-
3-ene. Phytochemistry 19, 2341–2346.
2 final
32P-labeled inserts of pVR1 cDNA were made using the
Prime-It II Random Labeling Kit (GibcoBRL). The 32P-
labeled probes were column-purified by elution through
Sephadex G 50.
4.12. Expression of isoflavanone reductase (Sor) activity in
E. coli
Recombinant SOR was produced and its activity mea-
sured as described above for the production of DMDI by
the Vr cDNA clone from alfalfa.
Matthews, D.E., Weiner, E.J., Matthews, P.S., VanEtten, H.D., 1987.
Role of oxygenases in pisatin biosynthesis and in the fungal degrada-
tion of maackiain. Plant Physiol. 83, 365–370.
Matthews, D.E., Plattner, R.D., VanEtten, H.D., 1989. The 6a-oxygen of
the pterocarpan glycinol is derived from molecular oxygen. Phyto-
chemistry 28, 113–115.
Acknowledgements
Miller, P.M., 1955. V-8 Juice agar as a general-purpose medium for fungi
and bacteria. Phytopathol. 45, 461–462.
Paiva, N., Edwards, R., Sun, Y., Hrazdina, G., Dixon, R., 1991. Stress
responses in alfalfa (Medicago sativa L.) 11. Molecular cloning and
expression of alfalfa isoflavone reductase, a key enzyme of isoflavonoid
phytoalexin biosynthesis. Plant Mol. Biol. 17, 653–667.
We thank Dr. Nancy Paiva, David Matthews, Qindong
Wu, and Yuegin Sun for technical advice and Dr. Cather-
ine Wasmann for editorial advice. This research was sup-
ported in part by Grant No. DE-FG03-00ER15078 from
the Department of Energy.
Paiva, N.L., Sun, Y., Dixon, R.A., VanEtten, H.D., Hrazdina, G., 1994.
Molecular cloning of isoflavone reductase from pea (Pisum sativum
L.): Evidence for
a 3R-isoflavanone intermediate in (+)-pisatin
biosynthesis. Arch. Biochem. Biophys. 312, 501–510.
References
Preisig, C.L., Matthews, D.E., VanEtten, H.D., 1989. Purification and
characterization of S adenosyl-L-methionine: 6a-hydroxymaackiain 3-
0-methyltransferase from Pisum sativum. Plant Physiol. 91, 559–566.
Preisig, C.L., Bell, J.N., Sun, Y., Hrazdina, G., Matthews, D.E.,
VanEtten, H.D., 1990. Biosynthesis of the phytoalexin pisatin:
isoflavone reduction and further metabolism of the product sophorol
by extracts of Pisum sativum. Plant Physiol. 94, 1444–1448.
Saghai-Maroof, M.A., Soliman, K.M., Jorgensen, R.A., Allard, R.W.,
1984. Ribosomal DNA spacer-length polymophisms in barley: men-
delian inheritance, chromosomal location, and population dynamics.
Proc. Natl. Acad. Sci. USA 81, 8014–8018.
Aoki, T., Akashi, T., Ayabe, S., 2000. Flavonoids of leguminous plants:
Structure, biological activity, and biosynthesis. J. Plant Res. 113, 475–
488.
Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G.,
Struhl, K., 1993. Current Protocols in Molecular Biology. John Wiley
and Sons, New York.
Banks, S.W., Dewick, P.M., 1982. Biosynthesis of the 6a-hydroxyptero-
carpan phytoalexin pisatin in Pisum sativum. Phytochemistry 21, 2235–
2242.
Banks, S.W., Dewick, P.M., 1983a. Biosynthesis of 6a-hydroxypterocar-
pans: deuterium NMR evidence for direct hydroxylation of pterocar-
pans. Z. Naturforsch. 38c, 185–188.
Banks, S.W., Dewick, P.M., 1983b. Biosynthesis of pisatin: experiments
with enantiomeric precursors. Phytochemistry 22, 1591–1595.
Cruickshank, I.A.M., Perrin, D.R., 1960. Isolation of a phytoalexin from
Pisum sativum L. Nature 187, 799–800.
Dewick, P.M., 1988. Isoflavonoids. In: Harborne, J.B. (Ed.), The
Flavonoids: Advances in Research Since 1980. Chapman & Hall,
London and New York, pp. 125–209.
Dewick, P.M., Ward, D., 1977. Stereochemistry of isoflavone reduction
during pterocarpan biosynthesis: an investigation using deuterium
nuclear magnetic resonance spectroscopy. J. Chem. Soc. Chem.
Commun., 338–339.
Shimada, N., Akashi, T., Aoki, T., Ayabe, S., 2000. Induction of
isoflavonoid pathway in the model legume Lotus japonicus: molecular
characterization of enzymes involved in phytoalexin biosynthesis.
Plant Sci. 160, 37–47.
Tiemann, K., Hinderer, W., Barz, W., 1987. Isolation of NADPH:iso-
flavone oxidoreductase, a new enzyme of pterocarpan phytoalexin
biosynthesis in cell suspension cultures of Cicer areitinum. FEBS Lett.
213, 324–328.
Tiemann, K., Dirk, I., Van Montagu, M., Barz, W., 1991. Pterocarpan
phytoalexin biosynthesis in elicitor-challenged chickpea (Cicer areiti-
num L.) cell cultures. Eur. J. Biochem. 200, 751–757.
Wu, Q., Preisig, C.L., VanEtten, H.D., 1997. Isolation of the cDNAs
encoding (+) 6a-hydroxymaackiain 3-O-methyltransferase, the termi-
nal step for the synthesis of the phytoalexin pisatin in Pisum sativum.
Plant Mol. Biol. 35, 551–560.
Dixon, R.A., 1999. Isoflavonoids: biochemistry, molecular biology, and
biological functions. In: Sankawa, U. (Ed.), Comprehensive Natural
Products Chemistry, Polyketides and Other Secondary Metabolites
Including Fatty Acids and Their Derivatives, vol. 1. Elsevier, Oxford,
pp. 773–823.
Wu, Q., VanEtten, H.D., 2004. Introduction of plant and fungal genes
into pea (Pisum sativum L.) hairy roots reduces their ability to produce
pisatin and affects their response to a fungal pathogen. Mol. Plant-
Microbe Interact. 17, 798–804.