Inactivation of pea genes by RNAi supports the involvement of two similar O-methyltransferases in the biosynthesis of (+)-pisatin and of chiral intermediates with a configuration opposite that found in (+)-pisatin

Article abstract of DOI:10.1016/j.phytochem.2007.06.013

(+)-Pisatin, the major phytoalexin of pea (Pisum sativum L.), is believed to be synthesized via two chiral intermediates, (-)-7,2′-dihydroxy-4′,5′-methylenedioxyisoflavanone [(-)-sophorol] and (-)-7,2′-dihydroxy-4′,5′-methylenedioxyisoflavanol [(-)-DMDI]; both have an opposite C-3 absolute configuration to that found at C-6a in (+)-pisatin. The expression of isoflavone reductase (IFR), which converts 7,2′-dihydroxy-4′,5′-methylenedioxyisoflavone (DMD) to (-)-sophorol, sophorol reductase (SOR), which converts (-)-sophorol to (-)-DMDI, and hydroxymaackiain-3-O-methyltransferase (HMM), believed to be the last step of (+)-pisatin biosynthesis, were inactivated by RNA-mediated genetic interference (RNAi) in pea hairy roots. Some hairy root lines containing RNAi constructs of IFR and SOR accumulated DMD or (-)-sophorol, respectively, and were deficient in (+)-pisatin biosynthesis supporting the involvement of chiral intermediates with a configuration opposite to that found in (+)-pisatin in the biosynthesis of (+)-pisatin. Pea proteins also converted (-)-DMDI to an achiral isoflavene suggesting that an isoflavene might be the intermediate through which the configuration is changed to that found in (+)-pisatin. Hairy roots containing RNAi constructs of HMM also were deficient in (+)-pisatin biosynthesis, but did not accumulate (+)-6a-hydroxymaackiain, the proposed precursor to (+)-pisatin. Instead, 2,7,4′-trihydroxyisoflavanone (TIF), daidzein, isoformononetin, and liquiritigenin accumulated. HMM has a high amino acid similarity to hydroxyisoflavanone-4′-O-methyltransferase (HI4′OMT), an enzyme that methylates TIF, an early intermediate in the isoflavonoid pathway. The accumulation of these four compounds is consistent with the blockage of the synthesis of (+)-pisatin at the HI4′OMT catalyzed step resulting in the accumulation of liquiritigenin and TIF and the diversion of the pathway to produce daidzein and isoformononetin, compounds not normally made by pea. Previous results have identified two highly similar "HMMs" in pea. The current results suggest that both of these O-methyltransferases are involved in (+)-pisatin biosynthesis and that one functions early in the pathway as HI4′OMT and the second acts at the terminal step of the pathway.

Full text of DOI:10.1016/j.phytochem.2007.06.013

Available online at www.sciencedirect.com
PHYTOCHEMISTRY
Phytochemistry 69 (2008) 76–87
www.elsevier.com/locate/phytochem
Inactivation of pea genes by RNAi supports the involvement
of two similar O-methyltransferases in the biosynthesis
of (+)-pisatin and of chiral intermediates with a configuration opposite
that found in (+)-pisatin
Evans Kaimoyo, Hans D. VanEtten *
Division of Plant Pathology and Microbiology, Department of Plant Sciences, 1140 E. South Campus Drive, Forbes 303,
University of Arizona, Tucson, AZ 85721, United States
Received 14 September 2006; received in revised form 12 May 2007
Available online 17 August 2007
Abstract
0
(
+)-Pisatin, the major phytoalexin of pea (Pisum sativum L.), is believed to be synthesized via two chiral intermediates, (ꢀ)-7,2 -dihy-
0
0
0
0
0
droxy-4 ,5 -methylenedioxyisoflavanone [(ꢀ)-sophorol] and (ꢀ)-7,2 -dihydroxy-4 ,5 -methylenedioxyisoflavanol [(ꢀ)-DMDI]; both
have an opposite C-3 absolute configuration to that found at C-6a in (+)-pisatin. The expression of isoflavone reductase (IFR), which
0
0
0
converts 7,2 -dihydroxy-4 ,5 -methylenedioxyisoflavone (DMD) to (ꢀ)-sophorol, sophorol reductase (SOR), which converts (ꢀ)-sopho-
rol to (ꢀ)-DMDI, and hydroxymaackiain-3-O-methyltransferase (HMM), believed to be the last step of (+)-pisatin biosynthesis, were
inactivated by RNA-mediated genetic interference (RNAi) in pea hairy roots. Some hairy root lines containing RNAi constructs of IFR
and SOR accumulated DMD or (ꢀ)-sophorol, respectively, and were deficient in (+)-pisatin biosynthesis supporting the involvement of
chiral intermediates with a configuration opposite to that found in (+)-pisatin in the biosynthesis of (+)-pisatin. Pea proteins also con-
verted (ꢀ)-DMDI to an achiral isoflavene suggesting that an isoflavene might be the intermediate through which the configuration is
changed to that found in (+)-pisatin. Hairy roots containing RNAi constructs of HMM also were deficient in (+)-pisatin biosynthesis,
but did not accumulate (+)-6a-hydroxymaackiain, the proposed precursor to (+)-pisatin. Instead, 2,7,4 -trihydroxyisoflavanone (TIF),
daidzein, isoformononetin, and liquiritigenin accumulated. HMM has a high amino acid similarity to hydroxyisoflavanone-4 -O-meth-
yltransferase (HI4 OMT), an enzyme that methylates TIF, an early intermediate in the isoflavonoid pathway. The accumulation of these
four compounds is consistent with the blockage of the synthesis of (+)-pisatin at the HI4 OMT catalyzed step resulting in the accumu-
0
0
0
0
lation of liquiritigenin and TIF and the diversion of the pathway to produce daidzein and isoformononetin, compounds not normally
made by pea. Previous results have identified two highly similar ‘‘HMMs’’ in pea. The current results suggest that both of these O-meth-
0
yltransferases are involved in (+)-pisatin biosynthesis and that one functions early in the pathway as HI4 OMT and the second acts at the
terminal step of the pathway.
ꢀ
2007 Elsevier Ltd. All rights reserved.
Keywords: Pisum sativum L.; Leguminosae; Pea; Biosynthesis; Isoflavonoids; Pterocarpans
1
. Introduction
+)-Pisatin 8 (see Fig. 1), a pterocarpanoid phytoalexin
alexin whose chemical structure was identified (Cruickshank
and Perrin, 1960; Perrin and Bottomley, 1962). Significant
progress towards elucidating the pathway for its biosynthe-
sis has been made through biochemical studies in pea and
related legumes (Dewick, 1988,1994; Dixon, 1999). The basic
pterocarpan carbon skeleton contains four rings ABCD as
shown for (+)-pisatin 8 in Fig. 1. Because of the presence
(
synthesized by pea (Pisum sativum L), was the first phyto-
*
Corresponding author. Fax: +1 520 621 7186.
E-mail address: vanetten@ag.arizona.edu (H.D. VanEtten).
0
031-9422/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2007.06.013

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
77
HO
O
4
2
3
7
5
'
'
formononetin 1
O
4
OCH3
HO
O
1
R
1
2
2
'-hydroxyformononetin R = H, R = OCH 2
O
3
HO
2
R
1
2
isoflavone reductase
DMD: R , R = OCH O 3
2
(IFR)
HO
O
HO
O
H
1
H
R
O
O
O
2
O
HO
HO
2
R
1
(
-)-vestitone: R = H, R = OCH
4
(
+)-sophorol 5a
3
1
2
(
-)-sophorol: R , R = OCH O 5
2
vestitone or sophorol
HO
O
reductase (VR or SOR)
HO
O
H
H
6
a
1
O
O
R
1
O
1a
H
H
(
+)-maackiain 6a
OH
HO
2
R
1
2
(
-)-DMI: R = H, R = OCH
9
HO
O
3
3
OH
1
2
(
-)-DMDI: R , R = OCH O 10
2
O
O
hydroxyisoflavanol
dehydratase
O
9
(HID)
(
+)-6a-hydroxymaackiain (6a-HMK) 7
HO
O
H
6
a-HMK methyltransferase
1
6
a
R
(
HMM)
CH O
O
B
11a
O
3
H
A
OH
2
R
O
O
1
2
(
-)-medicarpin: R = H, R = OCH 11
C
3
H
D
O
+
- isatin 8
1
2
(
-)-maackiain: R , R = OCH O 6
2
Fig. 1. Scheme for biosynthesis of the major pterocarpans of chick pea, alfalfa and pea. Solid arrows represent the steps in which enzymes have been
identified for the biosynthesis of (+)-pisatin 8 in pea, (ꢀ)-medicarpin 11 in alfalfa and (ꢀ)-maackiain 6 in chickpea, while dashed arrows represent a
proposal (Banks and Dewick, 1982a,b; Banks et al., 1983) for the biosynthesis of (+)-pisatin 8 through (+)-sophorol 5a, and (+)-6a-hydroxymaackiain 7.
The dotted line is the step(s) at which it is proposed intermediate(s) are formed so that the oxygen in the 6a hydroxy of (+)-pisatin 8 is derived from water.
0
0
0
0
0
0
0
0
DMD = 7,2 -dihydroxy-4 ,5 -methylenedioxyisoflavone 3, DMI = 7,2 -dihydroxy-4 -methoxy-isoflavanol 9, DMDI = 7,2 -dihydroxy-4 ,5 -methylenedi-
oxyisoflavanol 10.

78
E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
of sterogenic centers at the adjacent 6a and 11a positions in
pterocarpans, four stereoisomers are possible, but only two
forms, exemplified by (+)-6a and (ꢀ)-maackiain 6 (Fig. 1),
8. The biochemical basis for the conversion of (ꢀ)-DMDI
10 to (+)-6a-HMK 7, an intermediate with the opposite
configuration at C-6a, remains to be elucidated, but the
oxygen of the C-6a hydroxy group of (+)-pisatin 8 is
known to come from water (Matthews et al., 1987,1989).
Although data from early research supports the
involvement of the identified IFR, SOR and HMM pea
genes in (+)-pisatin 8 biosynthesis, pea also has the capac-
ity to synthesize (ꢀ)-maackiain 6 and (ꢀ)-medicarpin 11
derivatives (Stoessl, 1972; Pueppke and VanEtten, 1976)
and the identified IFR and SOR enzymes might be
involved only in the biosynthesis of these pterocarpans.
In the current study, RNA-mediated genetic interference
(RNAi) (Waterhouse et al., 1998,2001; Watson et al.,
2005) in transgenic pea hairy roots was used to directly
test the involvement of IFR, SOR and HMM in (+)-pisat-
in 8 biosynthesis. In addition, pea tissue was examined for
the presence of an enzymatic activity capable of convert-
ing (ꢀ)-DMDI 10 to precursors of (+)-6a-HMK 7. These
experiments were done with transgenic pea hairy roots
because of the rapidity and ease of producing transgenic
tissue compared to the time and effort required to produce
transgenic pea plants (Hamill and Lidgett, 1997; Wu and
VanEtten, 2004).
1
are known to occur in nature. While most legumes produce
pterocarpan phytoalexins with an absolute configuration as
shown
ꢀ)-maackiain 6 (Fig. 1), the garden pea is somewhat unu-
sual in that it synthesizes a (+)-pterocarpan, (+)-pisatin 8
for
(ꢀ)-medicarpin
11
and
(
(
Fig. 1) (Banks and Dewick, 1982b; Dewick, 1988).
In 1982, Banks and Dewick (1982a,b) suggested that the
biosynthesis of (+)-pisatin 8 proceeded via intermediates
such as (+)-sophorol 5a and (+)-maackiain 6a, which have
stereogenic centers with the same configuration as found in
(
+)-pisatin 8 (Fig. 1). However, recent studies have
0
suggested that the pathway, up to the (ꢀ)-7,2 -dihydroxy-
0
0
4
,5 -methylenedioxyisoflavanol [(ꢀ)-DMDI 10] intermedi-
ate, involve chiral intermediates with the opposite absolute
configuration at what will become carbon 6a in (+)-pisatin
8
(DiCenzo and VanEtten, 2006). Specifically, the
isoflavone reductase (IFR) catalyzing the conversion of
0 0 0
,2 -dihydroxy-4 ,5 -methylenedioxyisoflavone (DMD 3)
7
to (ꢀ)-sophorol 5 in pea (Fig. 1) has been characterized
and a cDNA for it isolated (Paiva et al., 1994). This
IFR, like those from alfalfa (Paiva et al., 1991), chickpea
(
Schlieper et al., 1990), and soybean (Fischer et al., 1990),
which make (ꢀ)-pterocarpans, converted the achiral iso-
flavone to a chiral (ꢀ)-isoflavanone with the configuration
at C-3 as shown for (ꢀ)-sophorol 5 (Fig. 1). There was no
evidence for the existence of an IFR that produces (+)-sop-
horol 5a from the achiral intermediate, DMD 3, or of an
epimerase that might produce (+)-sophorol 5a from
2. Results
2.1. Production of pea hairy root cultures
Wild type pea hairy roots (designated R1000ni) were
produced by infecting pea stem tissue with Agrobacterium
rhizogenes strain R1000 as described previously (Wu and
VanEtten, 2004). RNAi vectors consisting of two copies
of cDNA of IFR, SOR or HMM arranged as an inverted
repeat in the binary vector pFGC5941 were constructed.
The resulting vectors were used to produce hairy roots con-
taining the RNAi constructs. One hundred eighteen,
twenty-eight, and eight independent hairy root lines trans-
formed with the RNAi constructs with inverted cDNA
copies of HMM, SOR, and IFR, respectively, were
obtained. The yield of hairy root lines with the RNAi con-
structs of IFR, as measured by the number of hairy roots
obtained per total number of pea stems inoculated with
A. rhizogenes was low (4%) compared to that obtained with
the HMM RNAi constructs (65%) or the SOR RNAi con-
structs (18%).
(
ꢀ)-sophorol 5 in the protein extracts of (+)-pisatin-syn-
thesizing pea tissue (Paiva et al., 1991; DiCenzo and
VanEtten, 2006). In addition, precursor administration
studies using radiolabelled intermediates have demon-
strated that (ꢀ)-sophorol 5 is incorporated into (+)-pisatin
8
better than either (+)-sophorol 5a or (+)-maackiain 6a
(
Dicenzo and VanEtten, 2006). Furthermore, a sophorol
reductase, (SOR) that prefers (ꢀ)-sophorol 5 over (+)-sop-
horol 5a as a substrate, was identified in pea tissue synthe-
sizing (+)-pisatin 8 (DiCenzo and VanEtten, 2006). These
results support a (+)-pisatin 8 biosynthetic pathway that
proceeds through chiral intermediates with a configuration
opposite to that found in (+)-pisatin 8 (Fig. 1).
The current model for the late steps for (+)-pisatin 8
biosynthesis includes (ꢀ)-sophorol 5 and (ꢀ)-DMDI
1
0 as intermediates and has (+)-6a-hydroxymaackiain
[
(+)-6a-HMK 7] as the immediate precursor of (+)-pisatin
2.2. Time course analysis of (+)-pisatin biosynthesis in wild
type and RNAi pea hairy root lines
1
For simplicity, the configurations of pterocarpans are commonly
specified by their sign of rotation since the R/S designation is affected by
the influence of the substituents on the CIP-sequence rules. For example,
Pea hairy roots without an RNAi construct (R1000ni)
(+)-pisatin 8 is the (6aR, 11aR) isomer and (+)-maackiain 6a, which has
were treated with 0.08 mM CuCl to elicit (+)-pisatin 8
2
the same absolute configuration as (+)-pisatin 8, is (6aS, 11aS) isomer
because the presence of a H instead of a 6a-hydroxy changes the R/S
designation. Thus, in this paper, all isoflavonoids with the absolute
configurations shown for (+)-pisatin 8 will be labeled ‘+’ and those that
have the opposite will be labeled ‘ꢀ’.
production and the quantities produced were monitored
at 24 h intervals up to 72 h after elicitation. Non-elicited
wild type hairy root tissue, R1000ni, produced a mean of
5.6 lg/gFW of (+)-pisatin 8 (T = 0), and there was a

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
79
progressive increase to 77 lg/g FW by 72 h after elicitation
Fig. 2). The same response was obtained with control
hairy roots (R1000e) that contained the binary vector
pFGC5941 without inserts (data not shown). Transgenic
hairy roots silenced at each of the three enzymatic steps
contained little or no (+)-pisatin 8 at T = 0. When assayed
reduced amounts of (+)-pisatin 8, were compared to wild
type lines for (+)-pisatin 8 production over a 72-h period
(Fig. 2a). Transgenic hairy root lines HMMi #1, #3, #4,
#6 and #10 consistently showed no increase in (+)-pisatin
(
8 content up to 72 h after elicitation with CuCl (only data
2
for HMMi #10 are shown), while HMMi #55 showed a
slight increase in (+)-pisatin 8 (Fig. 2a) over the same
period.
2
4 h after elicitation, 23 of the 108 lines HMM RNAi con-
structs (HMMi lines) produced very little or no (+)-pisatin
(0–0.53 lg/g FW), 32 showed 50–80% reduction, and the
8
All of the eight transgenic hairy root lines carrying the
IFR RNAi construct to inactivate IFR expression (IFRi
lines) produced reduced amounts of (+)-pisatin 8, but none
was completely deficient in (+)-pisatin 8 biosynthesis, when
assayed 24 h after elicitation. Preliminary analysis of isofl-
avonoid production was completed but additional studies
were not done because the cultures did not survive long
enough (see below).
rest showed little (10–15%) or no reduction in (+)-pisatin 8
content when compared to non-transgenic hairy root line
R1000ni. Five of the transgenic lines, (HMMi #1, #3,
#4, #6, and #10) from the group that produced little or
no (+)-pisatin 8 were selected for further analysis. The
pisatin-deficient lines and transformant HMMi #55, repre-
senting the group of transgenic hairy roots that produced
Of the 28 independent lines containing the SOR RNAi
construct (SORi lines), five (SOR #2, #3, #6, #12, and
#
15) produced very little (<1 lg/g FW) or no (+)-pisatin
8 and four of them showed marked reduction in (+)-pisatin
production (4–15 lg/g FW) compared to controls at
a
8
6
4
2
0
0
0
0
8
T = 0 or at 24 h after elicitation. SORi #2 and SORi #15
were selected for further analysis and in the time-course
experiments, they showed very low amounts of (+)-pisatin
8
production 72 h after elicitation (Fig. 2b). SOR #21 is
representative of the group that produced a reduced
amount of (+)-pisatin 8. It demonstrated an intermediate
increase in the amounts of (+)-pisatin 8 over the 72 h per-
iod after elicitation.
0
0
24
48
72
Hrs after elicitation
2.3. HPLC profiles of secondary metabolites in wild type vs.
RNAi pea hairy roots
1
00
b
To determine whether the (+)-pisatin-deficient pheno-
type produced changes in the metabolic profiles of pea tis-
sue, chloroform-soluble metabolites extracted from both
wild type and RNAi hairy roots were analyzed by reversed
phase high performance liquid chromatography (HPLC).
When monitored at 309 nm, the HPLC profiles of second-
ary metabolites from wild type hairy roots (R1000ni) indi-
cated the presence of two major peaks. The main peak
corresponded to a compound that had a retention time
8
0
0
0
0
6
4
2
0
(R ) of 15.9 min and UV absorption spectrum of (+)-pisa-
t
0
24
48
72
tin 8 while the minor compound (R = 16 min) had the
t
Hrs after elicitation
properties of maackiain 6 (Fig. 3a). A subsequent analysis
for enantiomers demonstrated that these two compounds
were (+)-pisatin 8 and (ꢀ)-maackiain 6.
Fig. 2. Time-course analysis of the production of (+)-pisatin 8 in
transgenic vs. non-transgenic pea hairy root lines after elicitation with
.08 mM CuCl . (+)-Pisatin 8 was extracted by partitioning into hexanes
2
0
In contrast to the HPLC profile of isoflavonoid extracts
from R1000ni hairy roots (Fig. 3a), chloroform extracts
from HMMi #10 made 24 h after elicitation contained nei-
ther (+)-pisatin 8 nor (ꢀ)-maackiain 6 (Fig. 3b). The inac-
tivation of HMM expression by RNAi was expected to
lead to the accumulation of (+)-6a-HMK 7 (as predicted
in Fig. 1). However, this was not observed in transgenic
HMMi #10 or in any of the other independent lines (e.g.,
lines #1, #3, #4 and #6) that were silenced for HMM
expression. Instead, several compounds e.g., isoformo-
and the quantities of the phytoalexin were calculated from its absorbance
at 309 nm over a 72 h period. (a) (+)-Pisatin 8 produced by the control
hairy root line, R1000ni produced by infecting pea stem tissue with A.
rhizogenes (filled diamonds), transgenic hairy root lines containing RNAi
constructs of HMM, HMMi #10 (filled triangles) and HMMi #55 (filled
circles). (b) (+)-Pisatin 8 produced by control hairy root culture R1000ni
produced by infecting pea stem tissue with A. rhizogenes (filled diamonds),
transgenic hairy root cultures containing RNAi constructs of SOR, SORi
#
15 (open squares) and SORi #21 (open triangles). Error bars represent
means and standard deviations of three replicate samples. All experiments
were repeated 3 times with similar results.

80
E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
0
0
0
0
0
.080
.060
.040
.020
.000
a specific step in (+)-pisatin 8 biosynthesis. One of the IFRi
hairy root lines (IFRi #7) accumulated DMD 3 (Fig. 3c),
a
p
R = 10.5 min. DMD is the expected intermediate in the
t
conversion of DMD 3 to (ꢀ)-sophorol 5 by IFR (Fig. 1)
and its accumulation is consistent with the involvement
of the compound and IFR in (+)-pisatin 8 biosynthesis.
No further analyses of IFRi lines were carried out because
of the difficulty of obtaining and maintaining them.
HPLC profiles of secondary metabolites extracted from
SORi #2 (Fig. 3d) established the accumulation of (ꢀ)-sop-
m
th
b
0
.400
.300
.200
.100
dz
0
horol 5 (R = 8.45 min, 13 lg/gFW), as expected, and for-
l
if
t
mononetin 1 (R = 14.4 min, 1.1 lg/gFW), but little or no
t
0
0
hm
1
2
(
+)-pisatin 8 or (ꢀ)-maackiain 6. HPLC analysis of the
extracts of SORi transgenic hairy root lines #2 and #15,
as well as several others deficient in (+)-pisatin 8 biosynthe-
sis, also showed several major HPLC peaks including one
at Rt = 8 min with an UV absorption maximum at
348 nm. This compound also was observed in the extracts
of HMMi hairy root lines.
0
0
.000
.080
c
p
0
0
0
0
.060
.040
.020
.000
dm
2
.4. Metabolism of (ꢀ)-DMDI produced from (ꢀ)-sophorol
by protein extracts of pea
0
0
.015
.010
d
It was anticipated that RNAi silencing of the HMM-cat-
alyzed step would provide insight into the nature of the
intermediates that precede (+)-6a-HMK 7. However, the
lack of accumulation of the predicted intermediates raised
the possibility that the pathway was not silenced at the final
step (see below). Therefore, an alternative approach was
taken to identify the intermediate after (ꢀ)-DMDI 10, an
intermediate believed to precede (+)-6a-HMK 7 in the
(+)-pisatin 8 biosynthetic pathway. This alternative
approach sought to identify an enzymatic activity capable
of converting (ꢀ)-DMDI 10 to other possible intermediates
in the late steps of (+)-pisatin 8 biosynthesis.
s
f
0
0
.005
.000
2
4
6
8
10 12 14 16 18 20
Minutes
Fig. 3. HPLC profiles of secondary metabolites extracted from non-
transgenic hairy roots vs. those from hairy roots containing RNAi
contructs of HMM, IFR or SOR 24 h after elicitation with 0.08 mM
2
CuCl . (a) Non-transgenic control R1000ni, (b) transgenic HMMi #10, (c)
transgenic IFRi #7 and (d), transgenic SORi #2. p = (+)-pisatin 8, m =
To test whether (ꢀ)-DMDI 10 could be an intermediate
involved in (+)-pisatin 8 biosynthesis, (ꢀ)-DMDI 10 was
first synthesized from (ꢀ)-sophorol 5. Previous studies
0
(
ꢀ)-maackiain 6, th = 2,7,4 -trihydroxyisoflavanone 15, dz = daidzein 17,
l = liquiritigenin 14, if = isoformononetin 18, s = (ꢀ)-sophorol 5,
f = formononetin 1, hm = (+)-6a-hydroxymaackiain 7. Dashed arrow 1
in (b) represents the retention time for (+)-6a-HMK 7, while dashed arrow
(
W a¨ h a¨ l a¨ et al., 1997) have shown that the chemical reduc-
tion of isoflavones produce cis-10 and trans-isoflavan-4-ols
2
indicates the retention time of (+)-pisatin 8. Isoliquiritigenin and
1
0a (Fig. 4). The reduction of (ꢀ)-sophorol 5 by NaBH
4
isoformononetin 18 co-eluted at R = 13.9 min.
t
produced two major compounds (R = 6.34 and 7.0 min,
t
respectively (Fig. 5a), both of which have the molecular
mass of DMDI 10 (FW = 301) as empirically determined
by Atmospheric Pressure Chemical Ionization Mass Spec-
trometry (APCI-MS) (data not shown). When (ꢀ)-sopho-
rol 5 was incubated with SOR expressed in Escherichia
coli, the two DMDI 10 isomers also were produced with
nonetin 18 (R = 13.9 min, 26 lg/gFW), liquiritigenin 14
t
(
R = 6.6 min, 17.3 lg/gFW), daidzein 17 (R = 6.2 min,
t t
1
2 lg/gFW), isoliquiritigenin (R = 13.92 min, 14 lg/
t
0
gFW) and 2,7,4 -trihydroxyisoflavanone 15 (R = 5.5 min)
t
accumulated in the HMMi #10 (Fig. 3b). The other hairy
roots (HMMi #1, #3, #4 and #6) that also produced very
little or no (+)-pisatin 8 did not produce (+)-6a-hydrox-
ymaackiain 7 either but did produce significant amounts
of the same compounds identified in the extracts from
HMMi #10.
As indicated above, it was difficult to obtain a large
number of hairy root lines carrying the IFR RNAi con-
struct suggesting that silencing of this gene had pleiotropic
effects on the growth of hairy roots in addition to blocking
the major product being the isomer at R = 7 min. The
t
compound with R = 6.34 min has a maximum UV absorp-
t
tion at 300.5 nm with a shoulder at 285 nm (Fig. 5b) while
the one with R = 7 min has a maximum UV absorption at
t
299.4 nm with a shoulder at 285 nm (Fig. 5a and b). The
UV absorption spectra of both compounds were identical
to that previously reported by DiCenzo (1998) for the
(ꢀ)-DMDI 10 produced from (ꢀ)-sophorol 5 by the
recombinant pea SOR expressed in E. coli.

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
81
HO
O
6
HO
O
HO
O
a
OH
H
7
O
O
O
O
O
O
H
11a
4
O
HO
O
O
HO HO
HO
H
(
+)-6a-hydroxymaackiain 7
isoflavene (DMDIF) 12
O
O
cis -(-)-DMDI 10
+
+
HO
HO
O
HO
O
HO
O
(
-)-sophorol 5
7
OH
H
H
O
O
O
O
3
O
O
4
4
H
HO
HO
O
HO
2'
HO
(
-)-maackiain 6
3
,4,2'-triol 13
trans -(-)-DMDI 10a
Fig. 4. Proposed mechanism by which (+)-pisatin 8 is biosynthesized from intermediates with (ꢀ)-stereochemistry via an isoflavene intermediate. The
0
0
0
0
conversion of DMDIF 12 to 3,7,2 -trihydroxy-4 ,5 -methylenedioxyisoflavanol (3,4,2 triol 13) and then (+)-6a-hydroxypterocarpan, [(+) 6a-HMK 7] is
based on the intermediates made in the chemical synthesis of 6a-hydroxypterocarpans described by van Aardt et al. (2001).
0
0
0
0
.400
.350
.300
.250
0.050
.040
0.030
d (299.4 nm)
d (300.5 nm)
a
a
b
0
d
0
0
.020
.010
0
0
0
.200
.150
.100
0
0
.000
.080
0
0
.050
.000
i
0
0
0
.060
.040
.020
2
4
6
8
10 12 14 16 18
d
u
Minutes
0
.400
2
99.4
b
0
0
.350
.300
m
3
00.5
0.250
0.200
0.150
0.100
0.050
0.000
0
0
.000
.060
c
d
0
.040
s
0.020
0.000
2
40 260 280 300 320 340
nm
2
4
6
8
10 12 14 16 18 20
Minutes
Fig. 5. Production of (ꢀ)-DMDI 10 substrate. (ꢀ)-Sophorol 5 was
and products were analyzed by
HPLC. A two major peaks corresponding to the two isomers (cis and
trans) of (ꢀ)-DMDI 10, 10a. (b) UV absorption spectra corresponding to
the two major peaks in (a).
converted to (ꢀ)-DMDI 10 by NaBH
4
Fig. 6. Enzymatic conversion of (ꢀ)-DMDI 10 to an isoflavene by pea
crude protein extracts. (ꢀ)-DMDI 10 was purified by TLC and HPLC, (a),
and incubated with pea crude protein extracts for 2.5 h and the resulting
products were analyzed by HPLC. (b) HPLC trace for the analysis of
products obtained after the incubation of (ꢀ)-DMDI 10 with protein
extracts. (c) HPLC trace for the analysis of products obtained after the
Although, which of the two compounds was the cis-10
isomer and which was the trans-10a isomer of DMDI,
was not determined in this study, the isomer produced in
0
incubation of (ꢀ)-DMDI 10 with boiled pea protein extracts. d = 7,2 -
0
0
dihydroxy-4 ,5 -methylenedioxyisoflavanol 10, s = (ꢀ)-sophorol 5, m =
0
0
0
(
ꢀ)-maackiain 6, i = 7,2 -dihydroxy-4 ,5 -methylenedioxyisoflav-3-ene 12,
larger quantities (Fig. 5a, R = 7 min) by both the recombi-
t
u = unknown.
nant SOR expressed in E. coli and by NaBH -reduction of
4
(
ꢀ)-sophorol 5 was purified by thin layer chromatography
of DMDI 10 with pea crude protein extracts, the major
and HPLC (Fig. 6a). This was used as a substrate for pro-
tein extracts from elicited pea tissue. After 2.5 h incubation
compound produced was identified as an isoflavene,
0
0
0
7,2 -dihydroxy-4 ,5 -methylenedioxyisoflav-3-ene (DMDIF

82
E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
1
2), (R = 14.2 min) (Fig. 6b). Previous preliminary exper-
yltransferase resolved the mystery of how formononetin 1
is produced by legumes. The key initial steps for the bio-
synthesis of all isoflavonoids are summarized in Fig. 7
(Dixon, 1999; Aoki et al., 2000). It had long been thought
t
iments had also demonstrated the production of DMDIF
2 from (ꢀ)-DMDI 10 (Dicenzo, 1998). Maackiain (Rt
16.5 min) was also identified as a product (Fig. 6b) and
1
=
0
confirmed to be (ꢀ)-maackiain 6 by chiral HPLC analysis.
The other compounds produced could not be definitively
identified. The HPLC profiles indicated that the same com-
pounds were produced in both the absence and the pres-
ence of NADPH or FAD co-factors. A small amount of
the isoflavene 12 was produced with crude proteins
obtained from non-elicited tissue, but no DMDIF 12,
that methylation at the 4 position of daidzein 17 (dashed
arrow, Fig. 7) was responsible for the production of for-
mononetin 1. However, enzymes isolated from legumes
that methylated daidzein 17 always methylated the 7-
hydroxy group position of daidzein 17 (7IOMT), rather
0
than the 4 -hydroxy group, and produced isoformonone-
tin 18 (Liu and Dixon, 2001, and references within). A
gene for an enzyme, 2-hydroxyisoflavanone dehydratase
(HID), that dehydrates the 2,7-dihydroxyisoflavones has
been isolated from several legumes (Akashi et al., 2005;
Hakamatsuka et al., 1998) and has provided a better
understanding of the biosynthesis of both formononetin
1 and daidzein 17 by legumes (Fig. 7) as well as provided
insight into our results with the HMMi hairy root lines.
There are two copies of HMM in pea and two cDNAs of
HMM (HMM1 and HMM2) with deduced amino acid
sequences of 96% identity have been isolated from pea tis-
sue elicited to produce (+)-pisatin 8 (Wu et al., 1997). The
(
ꢀ)-maackiain 6 or any of the other unknowns were pro-
duced by boiled (10 min) protein extracts of elicited tissue
(
5
1
Fig. 6c). Non-metabolized DMDI 10 and (ꢀ)-sophorol
, presumably from the non-enzymatic oxidation of DMDI
0, and other unknowns were detected in the reaction mix-
ture containing the boiled proteins (Fig. 6c). Preliminary
tests indicated that DMDIF 12 could be metabolized fur-
ther by protein extracts from pea, but the products were
not identified (Kaimoyo, 2005).
0
3
. Discussion
HI4 OMT isolated from licorice shows a remarkably high
sequence similarity (>80% identity) to these HMMs and
0
When RNAi constructs of IFR, SOR, and HMM were
Akashi et al. (2003) speculated that the HI4 OMT of lico-
used to block the production of IFR, SOR and HMM pro-
teins, either no (+)-pisatin 8 or reduced levels of the phyto-
alexin were produced in most of the transgenic hairy root
rice may be functionally equivalent to HMM in pea.
Results of recent studies (Akashi et al., 2006) on the sub-
strate specificities of these two HMMs support Akasihi’s
hypothesis and are consistent with one of the HMMs
lines after elicitation with CuCl . In addition, the IFRi
2
0
and SORi hairy roots accumulated the two intermediates
(HMM1) being HI4 OMT and HMM2 being the specific
[
DMD 3 and (ꢀ)-sophorol 5, respectively] expected if
O-methyltransferase that methylates (+)-6a-HMK 7. The
0
IFR and SOR are involved in the biosynthesis of (+)-pisa-
tin 8. These results are consistent with previous results that
indicated that the synthesis of (+)-pisatin 8 involves some
of the same chiral intermediates that are involved in the
biosynthesis of (ꢀ)-pterocarpans (DiCenzo and VanEtten,
licorice HI4 OMT and HMM1 have 20- and 4-fold higher
0
methylation efficiencies (i.e., relative Vmax/K ) for 2,7,4 -
m
trihydroxyisoflavanone 15 than for (+)-6a-HMK 7. In con-
trast, HMM2 had a higher Vmax and lower Km on (+)-6a-
HMK 7 and methylated (+)-6a-HMK 7 67-fold more effi-
0
2
006; Paiva et al., 1994; Wu and VanEtten, 2004). Neither
ciently than it methylated 2,7,4 -trihydroxyisoflavanone 15.
DMD 3 nor (ꢀ)-sophorol 5 was detected in elicited or non-
elicited non-transgenic pea tissue, consistent with the
involvement of these intermediates in a channeled pathway
as lack of detection of intermediates is common for chan-
neled pathways (Hrazdina and Jensen, 1992; Winkel,
The two genomic copies of HMM in pea are within
18.7 kb or less of each other (Wu et al., 1997). Their close
proximity is consistent with a gene duplication of
0
HI4 OMT, a ubiquitous gene in legumes (Akashi et al.,
0
2003). The HI4 OMT could have retained its function as
2
004).
RNAi hairy roots blocked at HMM were expected to
in all other legumes while the duplicate could have evolved
in pea for the specific HMM activity. While the deduced
amino acid sequences of HMM1 and HMM2 are 96%
identical, modeling of the HMM proteins has identified
three or four active site residues that may be responsible
for their different substrate preferences (Akashi et al.,
2006). Because of their high sequence similarities, the pea
RNAi constructs of HMM (HMM1 was used) would be
expected to block expression of both HMM genes (Watson
et al., 2005). Thus, if one of the HMM genes in pea is the
accumulate (+)-6a-HMK 7 and to synthesize no (+)-pisa-
tin 8. Whereas some of the transgenic HMMi hairy root
lines were deficient in (+)-pisatin 8 biosynthesis, an accu-
mulation of (+)-6a-HMK 7 and other compounds that
might have provided insight into the late steps of (+)-
pisatin 8 biosynthesis, do not occur in any of the HMMi
hairy root lines (Fig. 3b). During the course of the current
study, another O-methyltransferase, hydroxyisoflavanone-
0
0
0
4
-O-methyltransferase (HI4 OMT), that catalyzes the
equivalent to the HI4 OMT found in most legumes, the
0
SAM-mediated methylation of 2,7,4 -trihydroxyisoflava-
none 15 to produce 2,7-dihydroxy-4 -methoxyisoflavanone
1
HMMi transgenic hairy roots of pea may not be able to
synthesize pisatin 8 (or maackiain 6) because the produc-
tion of formononetin 1 needed for the synthesis of these
compounds is blocked (Figs. 1 and 7). The blockage of for-
0
6 (Fig. 7) was characterized in licorice (Glycyrrhiza echi-
nata) (Akashi et al., 2003). The discovery of this O-meth-

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
83
OH
HO
O
C
7
HO
O
2
3
A
B
O
4'
OH
daidzein 17
O
2
-hydroxyiso-
liquiritigenin 14
flavanone dehydratase
2
-hydroxy-
(HID)
7-hydroxyisoflavone
methyltransferase
isoflavanone
synthase (IFS)
(7IOMT)
HO
O
O
OH
CH O
3
O
7
2
3
O
4'
OH
OH
isoformononetin 18
2,7,4'-trihydroxyisoflavanone 15
2,7,4'-trihydroxy-
isoflavanone 4'-
O-methyl
transferase (HI4'
O
MT)
HO
O
HO
O
O
OH
2-hydroxyiso-
flavanone dehydratase
O
(HID)
OCH3
formononetin 1
OCH3
2,7-dihydroxy-4'-methoxyisoflavanone 16
Fig. 7. Early known and hypothesized steps in the biosynthesis of isoflavonoids. The dashed arrow was the original proposed pathway by which
formononetin 1 was believed to be synthesized from daidzein 17.
mononetin 1 production could result in accumulation of
shown in Fig. 4. This triol would then be dehydrated by
an enzyme similar to the hydroxisoflavanol dehydratase
(HILD) that produces medicarpin 11 and maackiain 6
(Fig. 1). The combined effects would be the creation of a
new chirality with the additional feature that the oxygen
at the 6a-hydroxy position could be derived from water
because of the participation of the epoxide hydrolase.
Although Banks and Dewick (1983) did not find isofla-
venes to be incorporated well into (+)-pisatin 8, feeding
experiments in alfalfa (Martin and Dewick, 1978, 1980)
supported the participation of isoflavenes as precursors
to coumestans such as coumestrol. Coumestans are struc-
turally similar to pterocarpans but they have not been
reported in pea.
0
isoliquiritigenin, liquiritigenin 14, and 2,7,4 -trihydroxyi-
0
soflavanone 15 (Figs. 3b and 7). The 2,7,4 -trihydroxyisof-
lavanone 15 could be dehydrated by HID to produce
daidzein 17, which could subsequently be methylated by
1
7 7IOMT to give rise to isoformononetin 18 (Fig. 7).
The metabolism of the (ꢀ)-DMDI 10 isomer to the iso-
flavene DMDIF 12 (Fig. 4) by protein extracts from elic-
ited pea tissue (Fig. 6b) is of interest in that as an
intermediate, DMDIF 12 could both provide an explana-
tion of how the configuration is reversed in the (+)-pisatin
8
biosynthetic pathway as well as for the origin of the oxy-
gen of the C-6a hydroxy group. A recent chemical synthesis
0
of 6a-hydroxypterocarpans has an isoflavene and a 3,4,2
triol as intermediates 13 (van Aardt et al., 2001). The reac-
tion sequence can serve as a model for how the stereogenic
centers in (+)-pisatin 8 could be produced from the isofla-
vene DMDIF 12 and for how the oxygen in the 6a-hydroxy
group of pisatin 8 could be derived from water as demon-
strated by Matthews et al. (1987, 1989). In this scheme, an
epoxidase is hypothesized to catalyze the formation of an
epoxide at carbons 3 and 4 of the isoflavene, while an enan-
tioselective epoxide hydrolase (Botes et al., 2005) would
4. Concluding remarks
While studies on the production of phytoalexins by all
parts of the pea plant (Ingham, 1979) have consistently
shown that (+)-pisatin 8 is the only phytoalexin produced
in large amounts within 24–48 h after elicitation, Stoessl
(1972) reported that elicited pea pods synthesized small
amounts of (ꢀ)-maackiain 6. Thus, although it has long
0
subsequently convert this epoxide to the 3,4,2 triol 13

84
E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
been known that pea has the ability to synthesize (ꢀ)-
pterocarpans, the synthesis of pterocarpans with this con-
figuration is not commonly reported. In the current study,
the hairy roots that produced large amounts of (+)-pisatin
tin 8 using a strain of Aspergillus nidulans carrying the
pisatin demethylase gene (Preisig et al., 1989), and
(+)-pisatin 8 was extracted from pea tissue elicited for
(+)-pisatin 8 biosynthesis with copper chloride (Sweigard
et al., 1986). (ꢀ)-DMDI 10 was made from (ꢀ)-sophorol
5 by three different methods. In the first method,
(ꢀ)-DMDI 10 was made as described in DiCenzo and
VanEtten (2006) using an E. coli DH5a strain carrying a
cDNA (p2B1 or p3B1) for SOR. In the second method,
(ꢀ)-DMDI 10 was produced by the reduction of (ꢀ)-sop-
horol 5 with sodium borohydride as described by Anjaney-
ulu et al. (1965). (ꢀ)-Sophorol 5 (400 mg) EtOH (3 ml)
8
also produced small amounts of (ꢀ)-maackiain 6 (Fig. 3).
Thus, the ability of crude protein extracts from elicited pea
hairy root tissue to produce (ꢀ)-maackiain 6 from (ꢀ)-
DMDI 10 is consistent with the presence of a pea HILD
that produces a (ꢀ)-pterocarpan from a (ꢀ)-isoflavanol
as has been found in alfalfa and chickpea plants that are
synthesizing (ꢀ)-pterocarpans (Guo et al., 1994a,b). Of
particular significance with respect to the synthesis
of (+)-pisatin 8 in pea is the fact that the production of
were added to NaBH (6.7 mg) and shaken at 150 rpm
4
(
ꢀ)-maackiain 6 was also greatly reduced or blocked in
for 3 h at room temperature. The EtOH was evaporated
hairy roots that contained SORi (Fig. 3d vs. a). The block-
age of both (+)-pisatin 8 and (ꢀ)-maackiain 6 production
supports the hypothesis that in pea (ꢀ)-DMDI 10 is the
intermediate that serves as the branch point for the two dif-
ferent types of configuration (Fig. 1).
to dryness under a stream of nitrogen gas and H O
2
(2 ml) were added to the residue. (ꢀ)-DMDI 10 was
extracted by partitioning with CH Cl (4 ml) and purified
2
2
by thin layer chromatography (TLC) or HPLC.
(ꢀ)-DMDI 10 also was routinely made by bioconversion
of (ꢀ)-maackiain 6 using C. trifolii isolate T456. During the
conversion of (ꢀ)-maackiain to (ꢀ)-sophorol 5, the fungus
also converted part of the (ꢀ)-sophorol 5 to the same iso-
mer of (ꢀ)-DMDI 10 that is the major product made by
the recombinant SOR in E. coli. For the routine produc-
tion of (ꢀ)-DMDI 10 by this method, (ꢀ)-maackiain 6
(24 lg/ml) and 1% C. trifolii mycelium (w/v) in 500 ml of
GA broth (Matthews et al., 1987) were incubated with
shaking for 48 h at 25 ꢁC. With all procedures, care was
taken to avoid exposing (ꢀ)-sophorol 5 to basic conditions
to prevent racemization.
5
. Experimental
5
.1. Microorganisms and plasmid vectors
E. coli strain DH5a was used as the host for the binary
vectors and E. coli strain HB101, carrying the helper plas-
mid pRK2013 (Figurski and Helinski, 1979) was used to
mobilize the binary vectors from E. coli DH5a to A. rhiz-
ogenes during triparental mating. Plasmid pRK2013 carries
the nptII gene conferring kanamycin resistance. A. rhizoge-
nes strain R1000 (nalidixic acid resistant) was a gift from
Dr. M. Hawes, University of Arizona, Tucson, AZ. The
binary vector, pFGC5941was a gift from Dr. R. Jorgensen,
University of Arizona, Tucson, AZ: http://www.arabidop-
sis.org/abrc/catalog/vector_1.html. This vector is based on
binary vector pCAMBIA1200. It has a cassette bearing two
multiple cloning sites that facilitate the cloning of a gene in
the sense and anti-sense orientation relative to the CaMV
0
0
0
7,2 -Dihydroxy-4 ,5 -methylenedioxyisoflav-3-ene (DM-
DIF 12) was made from (ꢀ)-maackiain 6 by heating (ꢀ)-
maackiain 6 (2 mg) in concentrated HCl (100 ll) in EtOH
(3 ml) as described by Banks and Dewick (1982a). The
EtOH was dried and the resulting residue dissolved in
H O. The isoflavene was extracted from the aqueous phase
2
by partitioning with CH Cl (3· ml), which was reduced to
2
2
dryness under nitrogen gas, and the DMDIF 12 purified
from the residue by HPLC or TLC.
3
5S promoter thus producing inverted repeats of the
Formononetin 1 and daidzein 17 were purchased from
Sarsyntex (Merignac, France). Liquiritigenin 14 and isoli-
quiritigenin were purchased from Indofine Chemical Com-
pany (Hilsborough, NJ). Isoformononetin 18 was a gift
from Dr. R.A. Dixon, Samuel Roberts Noble Foundation,
inserted gene. The CaMV 35S promoter drives the tran-
scription of RNA inverted repeats. pFGC5941 also has a
copy of the basta resistance (bar) gene for the screening
of transgenic plants and the nptII gene for selection of
transformed bacteria.
0
Ardmore, OK and 2,7,4 -trihydroxyisoflavanone 15 was a
The fungus Colletotrichum trifolii isolate T456 is from a
previous study (Delserone et al., 1992) and p2B1 and p3B1,
two independent vectors carrying cDNA copies of SOR
cloned in pBluescript vector in E. coli DH5a, were from
previous work (DiCenzo and VanEtten, 2006.)
gift from Dr. S.-I. Ayabe, Nihon University, Japan. Tolu-
ene, EtOAc, MeOH and CH CN were purchased from
3
Honeywell International Inc. (Burdick and Jackson, Mus-
kegon, MI).
5.3. Construction of RNAi binary vectors for HMM, IFR
and SOR
5
.2. Chemicals
(
ꢀ)-Maackiain 6 was extracted from red clover roots
The HMM1 and SOR cDNAs were obtained from a
(
(
VanEtten et al., 1983), (ꢀ)-sophorol 5 was produced from
cDNA library constructed from RNA isolated from
infected pea seedlings (DiCenzo and VanEtten, 2006; Wu
et al., 1997). The pea IFR cDNA (Paiva et al., 1994) was
ꢀ)-maackiain 6 using C. trifoili isolate T456 (DiCenzo and
VanEtten, 2006), (+)-6a-HMK 7 was made from (+)-pisa-

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
85
a gift from Dr. N. Paiva, Samuel Roberts Noble Founda-
tion, Ardmore, OK. To construct the binary vectors for
days and were excised and transferred onto hormone- and
herbicide-free Gamborg’s B5 (Gamborg et al., 1968) semi-
solid gelriteꢂ medium supplemented with 3% sucrose and
0
0
RNAi, 400-bp fragments from the 5 - or 3 -end of each
cDNA of the three genes were amplified by PCR. The
PCR primers were designed to include BamHI and SwaI
ꢀ
1
500 mg l carbenicillin disodium salt. Hairy roots lacking
pFGC5941 were inhibited by basta allowing the resistance
0
ꢀ1
restriction sites at their 5 ends and XbaI and AscI sites
to basta (2 mg l ) to be used for selection of stable and clo-
0
at their 3 ends. HMM was amplified using HMM-3HF,
nal transgenic hairy root cultures. The integration of the
basta gene into the transformants was confirmed by RT-
0
5
-TCTAGAGGCGCGCCGGCACGAGTAGAAATG-
0
0
0
GA-3 and HMM-3HR, 5 -CGGGATCCA-TTTAAATC-
AACAACCTGTGGCTGATC-3 . SOR was amplified
with 5 -CCTCTAGAGGCGCGCCATGGCAGAGGG-
GAAAGG-3 and 5 -CGGGATCCATTTAA- ATGCAA-
CACTGAATCGC-3 . IFR was amplified with 5 -CCTCT
AGAGGCGCGCCATGGCAACTGAAAAC-3
0
PCR using the oligomers BarF, 5 -TGCACCAT-
0
0
0
CGTCAACCAC-3 and BarR, 5 -ACAGCGACCACG-
0
0
CTGTTGAA-3 . Hairy roots were sub-cultured every 3–4
0
0
weeks onto fresh Gamborg’s B5 plant tissue culture medium
semi-solidified with gelriteꢂ.
0
0
0
and
0
5
-CGGGATCCATTTAAATCTCATACACTGCATC-3 .
5.5. Elicitation, extraction and HPLC analyses of (+)-
pisatin and other metabolites
The PCR conditions for the amplification of each gene
were similar. They consisted of an initial heat denaturation
step at 95 ꢁC for 5 min, followed by 26 cycles of 94 ꢁC,
Hairy root tissue was elicited to produce (+)-pisatin 8
and other metabolites by immersing the tissue in
1
min, 58 ꢁC, 45 s, 72 ꢁC, 1 min, and a final 10 min cycle
at 72 ꢁC in a MJ Research PTC 100 thermal cycler (Global
Medical Instrumentation, Inc., Ramsey MN, USA).
Each PCR product was divided into two portions. The
first portion was digested with AscI and SwaI and cloned
into the AscI–SwaI restriction endonuclease sites of the
multiple cloning sites adjacent to the CaMV 35S promoter
of the binary vector pFGC5941. The resulting three binary
vectors contained one 400-bp fragment of the HMM, SOR
or IFR gene in the sense orientation relative to the CaMV
0.08 mM CuCl for 1 h at 24 ꢁC. The tissue was removed
2
from the CuCl solution, transferred to a beaker, and incu-
2
bated under high humidity in the dark for 24–72 h at 24 ꢁC.
In the initial screen for transgenic hairy roots deficient in
(+)-pisatin 8 biosynthesis and in the time course studies,
(+)-pisatin 8 was extracted by immersing the root tissue
in hexanes (6 v/g) and incubating the immersed roots on
a gyratory shaker at 130 RPM for 5–24 h. The hairy root
tissue was removed and the hexanes evaporated under vac-
uum. The residue was dissolved in EtOH (0.l–1 ml). The
presence of (+)-pisatin 8 in the extracts was determined
by scanning for the characteristic UV absorption spectrum
3
5S promoter. The second portion of each PCR amplifica-
tion product was digested with BamHI and XbaI and
cloned into the XbaI–BamHI sites of the second MCS of
pFGC5941. The three resulting RNAi vectors contained
two copies of the 400-bp fragment of HMM, SOR or
IFR arranged as an inverted repeat separated by the intron
carried by pFGC5941. The binary vectors were mobilized
into A. rhizogenes strain R1000 via the helper plasmid
pRK2013 using the triparental mating method of Deblaere
et al. (1987).
ꢃ
of (+)-pisatin 8 with a Beckman DU -64 UV absorbance
spectrophotometer (Beckman-Coulter, Fullerton, CA,
USA). The pisatin concentration was determined from its
absorbance at 309 nm using an extinction coefficient of
3
.86
10
(Cruickshank and Perrin, 1965).
To analyze transgenic hairy roots for the presence of
other secondary metabolites along with (+)-pisatin 8, hairy
root tissue was ground in 4 volume of EtOH–H O (7:3, v/
2
5
.4. Production of hairy roots
v). An additional 0.25 volume of H O were added to the
2
extract and the EtOH was evaporated in vacuuo. The aque-
ous phase was partitioned with EtOAc (4 vol), or CH Cl
Hairy roots of pea were constructed as previously
2
2
reported (Wen et al., 1999) with slight modifications. Briefly,
seeds of P. sativum L. var. Little Marvel (obtained from The
Meyer Seed Company, Baltimore, MD), were surface steril-
ized and grown on 0.8% water agar in magenta jars under
sterile conditions for 8–9 days. Short (10–12 mm) sections
of stem were cut and planted up side-down on MS (Murash-
(1 vol) or CHCl and the solvents were evaporated under
vacuum. The dry residue was dissolved in EtOH (0.1–
1 ml) ethanol and analyzed by HPLC.
3
Secondary metabolites were analyzed using an HPLC
system (Waters Inc. Milford, MA, USA) consisting of
two 515 pumps, a 717 plus Autosampler and a Photodi-
ode Array Detector 996. For most analyses, the metabo-
lites were separated using a C18 reversed phase column
(4 mm · 260 mm, Alltech or Discovery) and an
CH CN–H O gradient of 25–55% over a 20–60 min run
ꢀ1
ige and Skoog, 1962) medium semi-solidified with 2 g l
ꢀ1
gelriteꢂ and supplemented with 3% sucrose and 500 mg l
carbenicillin disodium salt. Each stem was inoculated with a
9
suspension culture (10–15 ll) of 1 · 10 CFU/ml A. rhizoge-
2
2
nes R1000 carrying the pFGC5941 with the inverted repeat
copies of SOR, HMM or IFR genes. A. rhizogenes carrying
no binary vector (R1000ni) or carrying pFGC5941 without
any inserted gene (R1000e) were used to generate hairy roots
serving as controls. Hairy roots began to emerge after 10
at the flow rates of 0.5–1 ml/min. To separate optical
isomers of pisatin 8 and maackiain 6, a chiral HPLC col-
umn (Chrialcel OD-RH, 4.6 · 150 mm) was used with a
CH CN–H O (1:1, v/v) solution run for 30 min. The
3
2
HPLC data were collected via an online system and ana-

86
E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
lyzed by Millennium 32ꢂ software. Authentic standards
were used to assist in the identification of the com-
pounds. TLC analysis was done on Analtech TLC plates
for his gift of isoformononetin both of the Samuel Roberts
Noble Foundation, Ardmore, OK, and Dr. S. Ayabe,
Nihon University, Japan for his gift of 2,7,4 -
0
(
UV detection at 254 nm; Uniplate, Newark, DE or EM
trihydroxyisoflavanone.
Sciences, silica gel 60 F254] Merck Kgaa, Darmstardt,
Germany) with solvent systems consisting of 10:1,
References
CHCl :MeOH or 1:1 and 60:40 toluene:EtOAc. After
3
purification by TLC, the metabolites were analyzed for
their UV absorption spectrum with a Beckman DU -64
UV absorbance spectrophotometer.
Some metabolites were further characterized by atmo-
spheric pressure chemical ionization (APCI) mass spec-
trometry and their m/z ratios compared to those of
authentic standards. The standards were pisatin 8 (m/z of
ꢃ
Akashi, T., Sawada, Y., Shimada, N., Sakurai, N., Aoki, T., Ayabe, S.,
2
003. cDNA cloning and biochemical characterization of S-adenosyl-
0
L-methionine: 2,7,4-trihydroxyisoflavanone 4 -O-methyltransferase, a
critical enzyme of legume isoflavonoid phytoalexin pathway. Plant
Cell. Physiol. 44, 103–112.
Akashi, T., Aoki, T., Ayabe, S., 2005. Molecular and biochemical
characterization of 2-hydroxyisoflavanone dehydratase. Involvement
of carboxylestrase-like proteins in Leguminous isoflavone biosynthesis.
Plant Physiol. 137, 882–891.
Akashi, T., VanEtten, H.D., Sawada, Y., Wasmann, C.C., Uchiyama, H.,
Ayabe, S., 2006. Catalytic specificity of pea O-methyltransferases
suggests gene duplication for (+)-pisatin biosynthesis. Phytochemistry
3
=
14), daidzein 17 (m/z = 254), isoformononetin 18 (m/z
0
268), 2,7,4 -trihydroxyisoflavanone 15 (m/z = 273),
0
0
0
(
ꢀ)-sophorol 5 (m/z = 300) and 7,2 -dihydroxy-4 ,5 -
methylenedioxyisoflav-3-ene 12 (m/z = 284).
6
7, 2525–2530.
Anjaneyulu, A.S.R., Sri Krishna, C., Row, R., 1965. Synthesis and study
of isoflavan-4-ols. Tetrahedron 21, 2677–2681.
Aoki, T., Akashi, T., Ayabe, S., 2000. Flavonoids of leguminous plants:
Structure, biological activity and biosynthesis. J. Plant Res. 113, 475–
5.6. In vitro metabolism of (ꢀ)-DMDI by protein extracts
of pea
4
88.
Elicited normal pea roots or hairy roots were incubated
at 23 ꢁC overnight and used immediately or frozen at
Banks, S.W., Dewick, P.M., 1982a. Biosynthesis of the 6a-hydroxyptero-
carpan phytoalexin pisatin in Pisum sativum. Phytochemistry 21, 2235–
2
Banks, S.W., Dewick, P.M., 1982b. (ꢀ)-Pisatin, an induced pterocarpan
metabolite of abnormal configuration from Pisum sativum. Phyto-
chemistry 21, 1605–1608.
ꢀ
80 ꢁC. Crude proteins were extracted from 2 g of elicited
242.
or non-elicited tissue by grinding the tissue to a fine pow-
der in liquid N and adding 4 ml Werth’s micro buffer
2
[
0.4 mM EDTA, 2 mM sodium bisulfite 7 mM sodium
Banks, S.W., Dewick, P.M., 1983. Biosynthesis of pisatin: experiments
with enantiomeric precursors. Phytochemistry 22, 1591–1595.
Botes, A.L., Lotter, J., Rhode, O.H.J., Botha, A., 2005. Interspecies
differences in the enantioselectivity of epoxide hydrolases in
Cryptococcus laurentii (Kufferath) C.E. Skinner and Cryptococcus
podzolicus (Bab’jeva & Reshetova) Golubev. Syst. Appl. Microbiol.
borate, 4 mM diethyl thiocarbamic acid, 50 mM ascorbic
acid, 10% (w/v) polyvinyl pyrolidone (PVP-40) and
0
.17% b-mecarptoethanol; pH 7.5] (Werth, 1985). The
suspension was centrifuged at 12,000 RPM for 15 min
at 4 ꢁC to obtain a supernatant containing the crude pro-
tein extract.
2
8, 27–33.
Cruickshank, I.A.M., Perrin, D.R., 1960. Isolation of a phytoalexin from
Pisum sativum L.. Nature 187, 799–800.
The crude protein extract (700 ll, 500–1200 lg) was
incubated with 50 ll of 0.3 mM (ꢀ)-DMDI 10 without
any co-factors or with 300 lM NADPH or FAD as co-fac-
tors at 26 ꢁC for 5, 15, 30 min, 2.5 h or overnight. As con-
trols, (ꢀ)-DMDI 10 was incubated with boiled (10 min)
crude protein extracts. The products of the enzyme reac-
tion were extracted by partitioning with 3 ml CH Cl
Cruickshank, I.A.M., Perrin, D.R., 1965. Studies on phytoalexins. VIII.
The effects of some further factors on the formation, stability and
localization of pisatin in vivo. Aust. J. Biol. Sci. 18, 817–828.
Deblaere, R., Reynaerts, A., H o¨ fte, H., Hernalsteens, J.-P., Leemans, J.,
Van Montagu, M., 1987. Vector for cloning in plant cells. Meth.
Enzymol. 153, 277–292.
Dewick, P.M., 1988. Isoflavonoids. In: Harborne, J.B. (Ed.), The
Flavonoids: Advances in Research Since 1980. Chapman & Hall,
London, pp. 125–209.
2
2
(
3 ml), with the CH Cl was removed by evaporation under
2 2
nitrogen. The residue was dissolved in EtOH and analyzed
by HPLC as described above.
Delserone, L.M., Matthews, D.E., VanEtten, H.D., 1992. Differential
toxicity of enantiomers of maackiain and pisatin to phytopathogenic
fungi. Phytochemistry 31, 3812–3819.
Dewick, P.M., 1994. Isoflavonoids. In: Harborne, J.B. (Ed.), The
Flavonoids. Affdvances in Research Since 1986. Chapman & Hall,
London, pp. 117–238.
Acknowledgements
DiCenzo, G.L., 1998. Study of the late steps of (+)-pisatin biosynthesis.
Ph.D. Dissertation, Department of Plant Pathology University of
Arizona.
DiCenzo, G.L., VanEtten, H.D., 2006. Studies on the late steps of (+)
pisatin biosynthesis: evidence for (ꢀ)-enantiomeric intermediates.
Phytochemisrty 7, 675–683.
Dixon, R.A., 1999. Isoflavonoid biochemistry, molecular biology and
biological functions. In: Sankawa, U. (Ed.), . In: Comprehensive
natural products chemistry. Polyketides and other secondary metab-
olites including fatty acids and other derivatives, vol. 1. Elsevier,
Oxford, pp. 773–823.
This work was a portion of a Ph.D. Dissertation by the
first author (Department of Plant Sciences, University of
Arizona, Tucson, AZ). The authors would like to thank
Dr. C. Wasmann for her editorial help and for her compar-
ative analysis of the HMM proteins, Dr. D. Gang and B.
Jackson for their help with APCI mass spectrometry anal-
ysis, F. Wen and Dr. M. Hawes for help in the production
of hairy roots, Dr. R. Jorgensen for pFGC5941, R.J.
Blount for his chiral HPLC analysis and Dr. R.A. Dixon,

E. Kaimoyo, H.D. VanEtten / Phytochemistry 69 (2008) 76–87
87
Figurski, D., Helinski, D.R., 1979. Replication of an original containing
derivative of plasmid RK2 dependent on a plasmid function provided
in trans. Proc. Natl. Acad. Sci. USA 76, 1648–1652.
Fischer, D., Ebeanau-Jehle, C., Grisebach, H., 1990. Phytoalexin synthesis
in soybean: purification and characterization of NADPH: 2-hydrox-
ydaidzein oxidoreductase from elicitor-challenged soybean cell culture.
Arch. Biochem. Biophys. 276, 390–395.
Gamborg, O.L., Miller, R.A., Ojima, K., 1968. Nutrient requirements of
suspension cultures of soybean root cells. Exp. Cell. Res. 50, 151–158.
Guo, L., Dixon, R.A., Paiva, N.L., 1994a. Conversion of vestitone to
medicarpin in alfalfa (Medicago sativa L.) is catalyzed by two
independent enzymes. Identification, purification, and characterization
and expression of alfalfa isoflavone reductase, a key enzyme of
isoflavonoid phytoalexin biosynthesis. Plant Mol. Biol. 17, 653–667.
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 3R-isoflavanone intermediate in (+)-pisatin biosyn-
thesis.. Arch. Biochem. Biophys. 312, 501–510.
Perrin, D.R., Bottomley, W., 1962. Studies on phytoalexins V. The structure
of pisatin from Pisum sativum L. J. Am. Chem. Soc. 84, 1919–1922.
Preisig, C.L., Matthews, D.E., VanEtten, H.D., 1989. Purification and
characterization of S-adenosyl-L-methionine: 6a-hydroxymaackiain-3-
O-methyltransferase from Pisum sativum. Plant Physiol. 91, 559–566.
Pueppke, S.G., VanEtten, H.D., 1976. Accumulation of pisatin and 3
additional antifungal pterocarpans in Fusarium solani-infected tissues
of Pisum sativum. Physiol. Plant Pathol. 8, 51–61.
0
0
of vestitone reductase and 7,2 -dihydroxy-4 -methoxyisoflavanone
dehydratase. J. Biol. Chem. 269, 22372–22378.
Guo, L., Dixon, R.A., Paiva, N.L., 1994b. The ‘‘pterocarpan synthase’’ of
alfalfa: Association and co-induction of vestitone reductase and 7,2-
dihydroxy-4-methoxy-isoflavanol (DMI) dehydratase, the two final
enzymes in medicarpin biosynthesis. FEBS Lett. 356, 221–225.
Hakamatsuka, T., Mori, K., Ishida, S., Ebizuka, Y., Sankawa, U., 1998.
Purification of 2-hydroxyisoflavone dehydratase from the cell cultures
of Pueraria lobata. Phytochemistry 49, 497–505.
Schlieper, D., Tiemann, K., Barz, W., 1990. Stereospecificity of hydrogen
transfer by fungal and plant NADPH: isoflavone oxidoreductases.
Phytochemistry 29, 1519–1524.
Stoessl, A., 1972. Intermin associated with pisatin in peas inoculated with
fungus Monilinia fructicola. Can. J. Biochem. 50, 107–108.
Sweigard, A.J., Matthews, D.E., VanEtten, H.D., 1986. Synthesis of the
phytoalexin pisatin by a methyltransferase from pea. Plant Physiol. 80,
277–279.
van Aardt, T.G., van Rensburg, H., Ferreira, D., 2001. Synthesis of
isoflavonoids. Enantiopure cis- and trans-6a-hydroxypterocarpans and
a racemic trans-pterocarpan. Tetrahedron 57, 7113–7126.
Hamill, J.D., Lidgett, A.J., 1997. Hairy root cultures: opportunities and
key protocols for studies in metabolic engineering. In: Doran, P.M.
(Ed.), Hairy Roots: Culture and Application. Harwood Academic
Publishers, Amsterdam, pp. 1–29.
Hrazdina, G., Jensen, R.A., 1992. Spatial-organization of enzymes in
plant metabolic pathways. Annu. Rev. Plant. Physiol. Plant Biol. 43,
VanEtten, H.D., Matthews, P.S., Mercer, E.H., 1983. (+)-Maackiain and
(+)-medicarpin as phytoalexins in Sophora japonica and identification of
the (ꢀ) isomers by biotransformation. Phytochemistry 22, 2291–2295.
W a¨ h a¨ l a¨ , K., Koskimies, J.K., Mesilaakso, M., Salakka, A.K., Leino,
T.K., Adlercreutz, H., 1997. The synthesis, structure and anticancer
activity of cis- and trans-4 ,7,-dihydroxyisoflavan-4-ols. J. Org. Chem.
62, 7690–7693.
Waterhouse, P.M., Graham, M.W., Wang, M.B., 1998. Virus resistance
and gene silencing in plants can be induced by simultaneous expression
of sense and antisense RNA. Proc. Natl. Acad. Sci. USA 95, 13959–
13964.
241–267.
Ingham, J.L., 1979. Isoflavonoid phytoalexins of yam bean (Pachyrrhizus
erosus). Z. Naturforsch. C 34, 683–688.
Kaimoyo E., 2005. Study of the (+)-pisatin biosynthetic pathway by
RNAi and development of a novel method to elicit the production of
plant secondary metabolites. Ph.D. Dissertation, University of Ari-
zona.
0
Liu, C.J., Dixon, R.A., 2001. Elicitor-induced association of isoflavone-O-
methyltransferase with endomembranes prevents the formation and
7
-O-methylation of daidzein during isoflavonoid phytoalexin biosyn-
Waterhouse, P., Wang, M.-B., Lough, T., 2001. Gene silencing as an
adaptive defense against viruses. Nature 411, 834–842.
Watson, J.M., Fusaro, A.F., Wang, M.-B., Waterhouse, P.M., 2005. RNA
silencing platforms in plants. FEBS Lett. 579, 5982–5987.
Wen, F., Zhu, Y., Hawes, M.C., 1999. Effects of pectin methyltransferase
gene expression on pea root development. Plant Cell 6, 1129–1140.
Werth, C.R., 1985. Implementing an isozyme lab at a field station.
Virginia J. Sci. 36, 53–76.
thesis. Plant Cell 13, 2643–2658.
Martin, M., Dewick, P.M., 1978. Role of an isoflav-3-ene in biosynthesis
of pterocarpan, isoflavan, and coumestan metabolites of Medicago
sativa. Tetrahedron Lett. 26, 2341–2344.
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.
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.
Winkel, B.S.J., 2004. Metabolic channeling in plants. Annu. Rev. Plant
Biol. 55, 85–107.
Wu, Q.D., 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
Microb. Int. 17, 798–804.
Murashige, T., Skoog, F., 1962. A revised medium for rapid growth and
bioassay with tobacco tissue cultures. Physiol. Plant. 15, 473–479.
Paiva, N.L., Edwards, R., Sun, Y., Hrazdina, G., Dixon, R.A., 1991.
Stress responses in alfalfa (Medicago sativa L.) 11. Molecular cloning
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.