Catalytic specificity of pea O-methyltransferases suggests gene duplication for (+)-pisatin biosynthesis

Article abstract of DOI:10.1016/j.phytochem.2006.09.010

S-adenosyl-l-methionine: 2-hydroxyisoflavanone 4′-O-methyltransferase (HI4′OMT) methylates 2,7, 4′-trihydroxyisoflavanone to produce formononetin, an essential intermediate in the synthesis of isoflavonoids with methoxy or methylenedioxy groups at carbon 4′ (isoflavone numbering). HI4′OMT is highly similar (83% amino acid identity) to (+)-6a-hydroxymaackiain 3-O-methyltransferase (HMM), which catalyzes the last step of (+)-pisatin biosynthesis in pea. Pea contains two linked copies of HMM with 96% amino acid identity. In this report, the catalytic activities of the licorice HI4′OMT protein and of extracts of Escherichia coli containing the pea HMM1 or HMM2 protein are compared on 2,7,4′-trihydroxyisoflavanone and enantiomers of 6a-hydroxymaackiain. All these enzymes produced radiolabelled 2,7-dihydroxy-4′-methoxyisoflavanone or (+)-pisatin from 2,7,4′-trihydroxyisoflavanone or (+)-6a-hydroxymaakiain when incubated with [methyl-¹⁴C]-S-adenosyl-l-methionine. No product was detected when (-)-6a-hydroxymaackiain was used as the substrate. HI4′OMT and HMM1 showed efficiencies (relative V_max/K_m) for the methylation of 2,7,4′-trihydroxyisoflavanone 20 and 4 times higher than for the methylation of (+)-6a-hydroxymaackiain, respectively. In contrast, HMM2 had a higher V_max and lower K_m on (+)-6a-hydroxymaackiain, and had a 67-fold higher efficiency for the methylation of (+)-6a-hydroxymaackiain than that for 2,7,4′-trihydroxyisoflavanone. Among the 15 sites at which HMM1 and HMM2 have different amino acid residues, 11 of the residues in HMM1 are the same as found in HI4′OMTs from three plant species. Modeling of the HMM proteins identified three or four putative active site residues responsible for their different substrate preferences. It is proposed that HMM1 is the pea HI4′OMT and that HMM2 evolved by the duplication of a gene encoding a general biosynthetic enzyme (HI4′OMT).

Full text of DOI:10.1016/j.phytochem.2006.09.010

PHYTOCHEMISTRY
Phytochemistry 67 (2006) 2525–2530
www.elsevier.com/locate/phytochem
Catalytic specificity of pea O-methyltransferases suggests
gene duplication for (+)-pisatin biosynthesis
a
b,
b
Tomoyoshi Akashi , Hans D. VanEtten *, Yuji Sawada a,1, Catherine C. Wasmann ,
a
a
Hiroshi Uchiyama , Shin-ichi Ayabe
a
Department of Applied Biological Sciences, Nihon University, Fujisawa, Kanagawa 252-8510, Japan
Division of Plant Pathology and Microbiology, Department of Plant Sciences, University of Arizona, Tucson, AZ 85721, USA
b
Received 31 May 2006; received in revised form 26 July 2006
Abstract
S-adenosyl-L-methionine: 2-hydroxyisoflavanone 40-O-methyltransferase (HI40OMT) methylates 2,7, 40-trihydroxyisoflavanone to
produce formononetin, an essential intermediate in the synthesis of isoflavonoids with methoxy or methylenedioxy groups at carbon
40 (isoflavone numbering). HI40OMT is highly similar (83% amino acid identity) to (+)-6a-hydroxymaackiain 3-O-methyltransferase
(HMM), which catalyzes the last step of (+)-pisatin biosynthesis in pea. Pea contains two linked copies of HMM with 96% amino acid
identity. In this report, the catalytic activities of the licorice HI40OMT protein and of extracts of Escherichia coli containing the pea
HMM1 or HMM2 protein are compared on 2,7,40-trihydroxyisoflavanone and enantiomers of 6a-hydroxymaackiain. All these enzymes
produced radiolabelled 2,7-dihydroxy-40-methoxyisoflavanone or (+)-pisatin from 2,7,40-trihydroxyisoflavanone or (+)-6a-hydrox-
ymaakiain when incubated with [methyl-14C]-S-adenosyl-L-methionine. No product was detected when (ꢀ)-6a-hydroxymaackiain was
used as the substrate. HI40OMT and HMM1 showed efficiencies (relative Vmax/Km) for the methylation of 2,7,40-trihydroxyisoflavanone
20 and 4 times higher than for the methylation of (+)-6a-hydroxymaackiain, respectively. In contrast, HMM2 had a higher Vmax and
lower Km on (+)-6a-hydroxymaackiain, and had a 67-fold higher efficiency for the methylation of (+)-6a-hydroxymaackiain than that
for 2,7,40-trihydroxyisoflavanone. Among the 15 sites at which HMM1 and HMM2 have different amino acid residues, 11 of the residues
in HMM1 are the same as found in HI40OMTs from three plant species. Modeling of the HMM proteins identified three or four putative
active site residues responsible for their different substrate preferences. It is proposed that HMM1 is the pea HI40OMT and that HMM2
evolved by the duplication of a gene encoding a general biosynthetic enzyme (HI40OMT).
Ó 2006 Elsevier Ltd. All rights reserved.
Keywords: Glycyrrhiza echinata; Pisum sativum; Leguminosae; Homology modeling; 2-Hydroxyisoflavanone; (+)-Pisatin; Pterocarpan
1. Introduction
niches (Aoki et al., 2000; Dixon, 1999). The first com-
pounds with an isoflavonoid skeleton in their biosynthesis
are 2,5,7,40-tetrahydroxyisoflavanone or 2,7,40-trihydroxyi-
soflavanone 2 (Fig. 1). These compounds, collectively
referred to as 2-hydroxyisoflavanone 2, are produced from
a general intermediate of flavonoid metabolism, (2S)-flava-
none 1, by the action of a legume-specific cytochrome P450
of the CYP93C subfamily (Akashi et al., 1999; Dixon,
1999). A subsequent intramolecular dehydration yields
isoflavones such as daidzein 5 (Akashi et al., 2005), which
lead to additional isoflavonoids such as the antimicrobial
pterocarpans (Dixon, 1999; Aoki et al., 2000).
Isoflavonoids are biosynthesized mainly by leguminous
plants and play important roles in the adaptation and
habituation of the producing plants to specific biological
Abbreviations: D7OMT, daidzein 7-O-methyltransferase; HI40OMT,
2,7,40-trihydroxyisoflavanone 40-O-methyltransferase; HMM, (+)-6a-hy-
droxymaackiain 3-O-methyltransferase; SAM, S-adenosyl-L-methionine.
*
Corresponding author. Tel.: +1 520 621 1977; fax: +1 520 621 7186.
E-mail address: vanetten@ag.arizona.edu (H.D. VanEtten).
Present address: RIKEN Plant Science Center, 1-7-22 Suehirocho,
1
Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan.
0031-9422/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2006.09.010

2526
T. Akashi et al. / Phytochemistry 67 (2006) 2525–2530
Fig. 1. Biosynthesis of isoflavones and the pterocarpan phytoalexins (+)-pisatin 9 in pea and (ꢀ)-medicarpin 7 in alfalfa and licorice. As shown, the pea
HMM isozymes are proposed to function at two different steps in the biosynthesis of pisatin. In this figure, only the 5-deoxyisoflavonoid pathway is
shown, but 5-hydroxyisoflavonoids are also produced by the IFS catalysis. The dotted arrows indicate that there are multiple steps between formononetin
4 and the indicated product. HMM, (+)-6a-hydroxymaackiain 3-O-methyltransferase; IFS, 2-hydroxyisoflavanone synthase; HI40OMT, 2,7,40-
trihydroxyisoflavanone 40-O-methyltransferase; HID, 2-hydroxyisoflavanone dehydratase; D7OMT, daidzein 7-O-methyltransferase.
Formononetin 4 (7-hydroxy-40-methoxyisoflavone) is
the precursor of the isoflavonoid derivatives with methoxy
or methylenedioxy groups at carbon 40 (isoflavone number-
ing) (Dixon, 1999). Formononetin 4 was proposed origi-
nally to be synthesized by the 40-O-methylation of
daidzein 5 (7,40-dihydroxyisoflavone), an isoflavone found
in soybean. However, the only known isoflavone O-methyl-
transferase in legumes that acts on daidzein 5 is the 7-O-
methyltransferase (D7OMT) that yields isoformononetin
6 (7-methoxy-40-hydroxyisoflavone) (Akashi et al., 2003).
It is now known that formononetin 4 is produced by the
action of 2,7,40-trihydroxyisoflavanone 40-O-methyltrans-
ferase (HI40OMT), first characterized in licorice (Glycyr-
rhiza echinata L.) (Akashi et al., 2003). HI40OMT acts on
2-hydroxyisoflavanone 2 to produce 2,7-dihydroxy-40-
methoxyisoflavanone 3, which is dehydrated to yield for-
mononetin 4 (Akashi et al., 2005). D7OMT and HI40OMT
have only about 50% identity and no substrate in common.
On the other hand, HI40OMT has 83% identity with (+)-
6a-hydroxymaackiain 3-O-methyltransferase (HMM) of
pea (Pisum sativum L.) (Wu et al., 1997). HMM catalyzes
the last step in the biosynthesis of (+)-pisatin 9, a pterocar-
panoid phytoalexin, which is unusual in having a (+)-chi-
rality (Fig. 1). Most pterocarpan phytoalexins, such as
(ꢀ)-medicarpin 7 found in alfalfa and licorice, have a
(ꢀ)-chirality (Dewick, 1988).
In this report, the catalytic activities of a purified recombi-
nant licorice HI40OMT protein and of extracts of E. coli
containing the pea HMM1 or HMM2 protein are com-
pared on 2,7,40-trihydroxyisoflavanone 2, as well as on
enantiomers of 6a-hydroxymaackiain 8a and 8b. The
results indicate that HI40OMT and HMM1 have very sim-
ilar activities but that HMM1 and HMM2 differ greatly in
their substrate preferences. These results combined with
homology modeling data have lead us to a proposal for
the molecular evolution of the specific O-methyltransferase
involved in the biosynthesis of (+)-pisatin 9.
2. Results and discussion
Both HI40OMT and the HMMs produced the expected
radiolabelled products when [methyl-14C]-S-adenosyl-L-
methionine (SAM) along with 2,7,40-trihydroxyisoflava-
none 2 or (+)-6a-hydroxymaakiain 8a were used as the
substrates. No product was detected when (ꢀ)-6a-hydrox-
ymaackiain 8b was used as the substrate.
The steady-state kinetic data for HI40OMT and the
HMM enzymes are shown in Table 1. The Michaelis con-
stants (Km’s) for the reactions of HI40OMT and HMM1
with the two methyl acceptors were 3 lM for 2,7,40-trihydr-
oxyisoflavanone 2 and slightly higher (6–8 lM) for (+)-6a-
hydroxymaackiain 8a, indicating that these proteins have
somewhat greater affinity for 2,7,40-trihydroxyisoflavanone
2 than for (+)-6a-hydroxymaackiain 8a. In contrast,
HMM2 showed a higher Km (23 lM) for 2,7,40-trihydr-
oxyisoflavanone 2 and a lower Km (0.5 lM) for (+)-6a-
hydroxymaackiain 8a indicating an approximately 50-fold
higher affinity for the pisatin 9 precursor than for 2,7,40-tri-
hydroxyisoflavanone 2. The Vmax and Vmax/Km values can-
not be compared because the HMMs were not purified
proteins. However, one can compare the efficiencies of
In pea, two closely linked HMM genes are found, and
the HMM cDNAs encode the isozymes, HMM1 and
HMM2, which share 96% amino acid sequence identity
(Wu et al., 1997). Of the substrates tested, both HMMs
showed highest activity on (+)-6a-hydroxymaackiain 8a
and essentially no activity on (ꢀ)-6a-hydroxymaackiain
8b (structure not shown). Interestingly, while the licorice
HI40OMT preferred 2-hydroxyisoflavanone 2 as a sub-
strate, it also had activity on ( )-medicarpin suggesting
substrate overlap with the HMMs (Akashi et al., 2003).

T. Akashi et al. / Phytochemistry 67 (2006) 2525–2530
2527
Table 1
Kinetic properties of recombinant licorice HI40OMT, pea HMM1 and HMM2
Substrate
Km [lM]
Vmax [pkatal mgꢀ1
]
Vmax/Km [pkatal mgꢀ1 lMꢀ1
]
Vmax/Km [%]c
HI40OMTa
HMM1b
HMM2b
2,7,40-Trihydroxyisoflavanone 2
(+)-6a-Hydroxymaackiain 8a
3
8
20,000
2700
6600
340
100
5
2,7,40-Trihydroxyisoflavanone 2
(+)-6a-Hydroxymaackiain 8a
3
6
1800
910
600
150
100
25
2,7,40-Trihydroxyisoflavanone 2
(+)-6a-Hydroxymaackiain 8a
23
0.5
340
520
15
1000
100
6700
The values are the averages from two independent experiments (maximum deviation was about 10%).
a
Purified recombinant licorice HI40OMT.
10,000g supernatant of an extract of Escherichia coli expressing pea HMM1 or HMM2.
Relative ratio of Vmax/Km within the individual enzyme reaction to the two methyl acceptors (2,7,40-trihydroxyisoflavanone = 100%).
b
c
the enzyme reactions in terms of relative Vmax/Km with
each substrate (see Table 1). Whereas the Vmax ratios of
HI40OMT and HMM1 with (+)-6a-hydroxymaackiain 8a
were <50% of their ratios with 2,7,40-trihydroxyisoflava-
none 2, HMM2 had a 1.5-fold higher Vmax on (+)-6a-
hydroxymaackiain 8a than on 2,7,40-trihydroxyisoflava-
none 2. As a result, HI40OMT and HMM1 showed
efficiencies (relativeVmax/Km) for the methylation of 2,7,40-
trihydroxyisoflavanone 2 20 and 4 times higher than the
efficiencies for (+)-6a-hydroxymaackiain 8a, respectively.
In contrast, HMM2 showed a 67-fold higher efficiency
for the methylation of (+)-6a-hydroxymaackiain 8a than
that for 2,7,40-trihydroxyisoflavanone 2. Thus, we suggest
that in pea, HMM1 is the equivalent of the HI40OMT
Fig. 2. Alignments (a) and a phylogenetic tree (b) of the HMMs of pea and related OMT proteins. (a) The aligned amino acid sequences of HMM2,
HMM1, the HI40OMTs of licorice (Ge40OMT) and Lotus japonicus (Lj40OMT), and a putative HI40OMT of Medicago truncatula (Mt40OMT). Dots
indicate the amino acid residues identical to those of HMM2. Gaps (-) are inserted to optimize alignment. The conserved catalytic His 264 is denoted by
the blue box, and the residues that differ between HMM1 and HMM2 in the putative active site are indicated with red boxes. Amino acid residues different
between HMM1 and HMM2 are marked by stars: blue, unique amino acid residues in HMM2; green, residues common to HMM2 and the HI40OMTs. (b)
The tree was constructed from the amino acid sequences of the proteins in (a) with the alfalfa D7OMT as the outgroup. The numbers above the branches
are the bootstrap values for 1000 replicates. The scale bar shows the amino acid substitution ratio.

2528
T. Akashi et al. / Phytochemistry 67 (2006) 2525–2530
found in other legumes and is involved in formononetin 4
biosynthesis and that HMM2 likely has a different sub-
strate in planta and is involved primarily in the biosynthesis
of (+)-pisatin 9.
The hypothesized binding of the substrates to the com-
mon active sites of HI40OMT/HMMs was verified first in
silico. When plausible stereostructures of (+)-6a-hydrox-
ymaackiain 8a and the four optical isomers of 2,7,40-tri-
Fig. 2a shows the alignment of the amino acid sequences
of the HMMs and HI40OMTs of licorice, Lotus japonicus,
and Medicago truncatula. In total, there are 15 sites at which
HMM1 and HMM2 have different amino acid residues
(indicated with stars over the sequences in Fig. 2a). At 11
(blue stars) of these 15 sites, the amino acid residues in
HMM1 and the other HI40OMTs are identical. The amino
acid residues at these 11 sites are not conserved in HMM2.
Therefore, these amino acid residues could be important in
determining the differential substrate preferences of the
HMM1/HI40OMTs and HMM2. The structure of the
alfalfa D7OMT protein, based on X-ray crystallography,
has been reported by Zubieta et al. (2001). Also, the homol-
ogy models of chavicol O-methyltransferase and eugenol O-
methyltransferase of sweet basil based on the alfalfa
D7OMT structure have identified the amino acid residues
that are critical for the substrate preferences of these OMTs
(Gang et al., 2002). The amino acid identity of the chavicol/
eugenol O-methyltransferases and D7OMT is ꢁ43%,
whereas the identity of HMM/HI40OMT and D7OMT is
ꢁ49%. Therefore, the D7OMT-based homology models
of the HMM proteins are expected to be more useful for
identifying which of the unconserved amino acid residues
are critical in determining the substrate preferences of the
HMM isozymes.
hydroxyisoflavanone
2
are compared, the overall
structures of the (2R, 3S)- and (2S, 3R)-forms with their
quasi-equatorial 2-hydroxy and 3-phenyl conformations
can be superimposed onto (+)-6a-hydroxymaackiain 8a
(Fig. 3a). We favor the (2R, 3S)-2,7,40-trihydroxyisoflava-
none 2 as the likely substrate as we expect it to be the initial
product from the action of the CYP93C involved in the syn-
thesis of 2-hydroxyisoflavanone 2 from the (2S)-flavanone 1
(Fig. 1) (Sawada et al., 2002; Sawada and Ayabe, 2005). The
homology modeled structures of the HMM1 and HMM2
proteins also had the same overall conformation (data not
shown). Among the 15 sites where the residues of HMM1
and HMM2 differ, four sites (residues 120, 128, 132 and
167) were located at the substrate-binding region (indicated
with red boxes in the alignment (Fig. 2a)). While L120/V120
of HMM1/HMM2 both differ from the corresponding res-
idue (I) of HI40OMT, S128, M132 and S167 are conserved
in HMM1 and HI40OMT. A closer inspection of the active
site (Fig. 3b) shows that when the hydroxyls of the sub-
strates are positioned next to the conserved catalytic H264
(blue) involved in their methylation, the different hydropho-
bic residues, M132 (red) of HMM1 and the L132 (green) of
HMM2, are near the 7-hydroxyl of 2-hydroxyisoflavanone
2 and the methylenedioxy of (+)-6a-hydroxymaackiain 8a.
Also, the differences at positions 128 (S vs. I) and 167 (S
Fig. 3. Stereoviews of the substrates of HI40OMT and HMM (a) and the active site of a homology-modeled HMM with the substrates superimposed and
docked (b). (a) Substrates are shown as balls and sticks. Carbons of (+)-6a-hydroxymaackiain 8a, (2R,3S)- and (2S,3R)-2,7,40-trihydroxyisoflavanone 2
are displayed as pink, yellow and gray balls, respectively. Oxygens are shown as red balls, and hydrogens at C-11a of (+)-6a-hydroxymaackiain 8a and C-2
and C-3 of 2,7,40-trihydroxyisoflavanones 2 are displayed as white balls. (b) Conserved catalytic H264 (blue), the amino acid residues at positions 120, 128,
132 and 167 of HMM1 (red) and HMM2 (green), (+)-6a-hydroxymaackiain 8a [(+)-6a-HM, pink] and (2R,3S)-2,7,40-trihydroxyisoflavanone 2 [(2R,3S)-
˚
2HI, yellow] are shown as sticks. The distance between H264 and each of the methyl-accepting hydroxyls of the substrates is about 3 A. Residues at
˚
˚
position 132 (M and L) are located within 4 A, and residues at positions 128 (S and I) and 167 (S and Y) are within 8 A, of the 7-hydroxyl of 2-
hydroxyisoflavanone 2 and the methylenedioxy of (+)-6a-hydroxymaackiain 8a, respectively.

T. Akashi et al. / Phytochemistry 67 (2006) 2525–2530
2529
vs. Y) involve amino acids which differ in size, hydrophobic-
ity and in the presence of a hydroxyl, which if present, might
participate in hydrogen bonding with the substrate (posi-
tion 128). We expect the residues at these three sites, and
possibly at position 120 as well, to be involved in determin-
ing the differences in the catalytic activities of HMM1 and
HMM2, particularly the high affinity (low Km) of HMM2
for the (+)-pisatin 9 precursor.
duplication of a gene encoding a general biosynthetic
enzyme (HI40OMT) resulted in the emergence of a second
enzyme that acts later in the same pathway for the produc-
tion of a unique phytoalexin, (+)-pisatin 9.
4. Experimental
HI40OMT is expected to be a highly conserved gene
found in all legumes that produce isoflavonoids with meth-
oxy and methylenedioxy groups at the carbon equivalent to
the 40-position of an isoflavone. Pea is one such legume and
the enzymatic data presented here indicate that what was
originally identified as HMM1 is the pea HI40OMT. In
contrast, HMM2, which has a very high efficiency for the
methylation of (+)-6a-hydroxymaackiain 8a, would appear
to be an enzyme primarily involved in the last step of (+)-
pisatin 9 biosynthesis. These results are consistent with an
interesting mechanism for the molecular evolution of
enzymes of secondary metabolism. It has been demon-
strated that two (or more) HMM genes, presumably
HMM1 and HMM2, are present on an 18.7-kb BamHI
fragment of the pea genome (Wu et al., 1997). Several
examples of clusters of paralogous genes encoding enzymes
of secondary metabolism with different specificities are
found in plants (Shimada et al., 2003; Shimada et al.,
2005), and gene duplication followed by the accumulation
of mutations is an accepted mechanism to account for the
extreme diversity of plant secondary products (Ober, 2005).
The phylogenetic tree indicates that the pea HMMs
form a monophyletic clade with strong bootstrap support
in a clade of legume HI40OMTs (Fig. 2b). The branch
length of HMM2 (0.033) is about five times longer than
that of HMM1 (0.007), suggesting a higher accumulation
of amino acid substitutions in the HMM2 sequence than
in the HMM1 sequence. Also, a reconstruction of the
ancestral sequence of the HMMs revealed that 13 amino
acid residues (V24I, M37V, H80Y, E93G, L120V, S128I,
M132L, C167Y, M170I, G244S, N277K, E300D,
R305H), including all the active site residues highlighted
in Fig. 2a and Fig. 3b, have been replaced to generate
the HMM2 sequence in contrast to only three amino acids
(A58S, E94G, C167S) for the HMM1 sequence.
4.1. Chemicals
The (+)- and (ꢀ)-6a-hydroxymaackiain 8a and 8b were
obtained as previously described (VanEtten et al., 1983).
The 2,7,40-trihydroxyisoflavanone 2 was prepared in vitro
using yeast microsomes expressing CYP93C2 (Ayabe
et al., 2002).
4.2. Enzymes and enzyme assay
To prepare the HMM isozymes, HMM1 and HMM2
cDNAs in pBluescript SK(ꢀ) were transformed into
E. coli DH5a. Extracts containing HMM1 and HMM2
were prepared as described previously (Wu et al., 1997)
except the induction by b-D-thiogalactopyranoside was
for 5 h and the extraction buffer was 100 mM potassium
phosphate buffer (pH 7.5) containing 10% (w/v) sucrose
and 14 mM 2-mercaptoethanol. Purified recombinant lico-
rice HI40OMT with six histidine residues at the N-terminus
was prepared as previously described (Akashi et al., 2003).
HI40OMT activity was assayed in a reaction mixture
containing 2,7,40-trihydroxyisoflavanone 2 (Akashi et al.,
2003) dissolved in 2 ll 2-methoxyethanol, 150 lM SAM
(2.4 nmol [14C]SAM (2.26 GBq/mmol, Amersham Biosci-
ences, Buckinghamshire, UK) and 12.6 nmol unlabeled
SAM) and either the crude extract of E. coli with HMM1
(350 ng protein), or HMM2 (1.75 lg protein) or the puri-
fied recombinant HI40OMT (17.5 ng protein), in a total
volume of 100 ll. To determine the kinetic parameters, 1,
3, 9, 30 and 100 lM of 2,7,40-trihydroxyisoflavanone 2
were used. The reaction was carried out at 30 °C for
5 min and terminated by the addition of EtOAc (60 ll).
The EtOAc extract of the mixture (30 ll) was subjected
to silica-gel TLC [LK6DF (Whatman, Maidstone, UK)]
using CHCl3:acetone:25% aq. ammonia solution (70:29:1,
v/v). The product, [14C]-2,7-dihydroxy-40-methoxyisoflava-
none 3, was collected from the developed TLC and its
radioactivity measured by liquid scintillation counting.
Km and Vmax values were calculated using a Lineweaver–
Burk plot. For the assay of HMM activity, (+)-6a-hydrox-
ymaackiain 8a (0.3, 0.75, 1.5, 3, 9 or 30 lM) was incubated
with E. coli extracts containing HMM1 (875 ng protein), or
HMM2 (1.17 lg protein) or the purified recombinant
HI40OMT (350 ng) in the presence of 150 lM SAM
(2.4 nmol [14C]SAM and 12.6 nmol unlabeled SAM) in a
total volume of 100 ll at 30 °C for 5 min (Preisig et al.,
1989; Wu et al., 1997). For resolving the 6a-hydrox-
ymaackiain O-methyltransferase product, LK6DF TLC
plates were developed with toluene:EtOAc (3:2, v/v).
3. Concluding remarks
A consideration of the amino acid substitutions leading
to HMM2 and the molecular modeling studies has allowed
the identification of four active site residues predicted to be
critical for the differences in HMM1 and HMM2 catalysis.
While most legumes produce pterocarpans with (ꢀ)-chiral-
ity such as (ꢀ)-medicarpin 7 (Fig. 1), the garden pea is
somewhat unusual in that it synthesizes a (+)-pterocarpan,
(+)-pisatin 9 (Dewick, 1988). Thus, it would be expected
that there would be specific enzymes involved in the synthe-
sis of this pterocarpan. It is hypothesized that in pea, the

2530
T. Akashi et al. / Phytochemistry 67 (2006) 2525–2530
L-methionine: 2,7,40-trihydroxyisoflavanone 40-O-methyltransferase, a
4.3. Homology modeling
critical enzyme of the legume isoflavonoid phytoalexin pathway. Plant
Cell Physiol. 44, 103–112.
The structural formulae of 2,7,40-trihydroxyisoflava-
none 2 and (+)-6a-hydroxymaackiain 8a were drawn using
ChemDraw Pro (CambridgeSoft Corporation, Cambridge,
MA, USA). The stereochemical analysis was performed
using DS ViewerPro 5.0 (Accelrys Inc., San Diego, CA,
USA). The three-dimensional homology models of
HMM1 and HMM2 were constructed using MODBASE
program (http://salilab.org/modbase). Of the 10 HMM1
and HMM2 models constructed, only those, which aligned
with the reported stereostructure of D7OMT (Zubieta
et al., 2001), were selected for further analyses. The super-
imposition of the HMM1 and HMM2 models was carried
out using i-Mol software (http://www.pirx.com/iMol/).
The substrate–protein complex was generated by the man-
ual docking of the individual structures using DS Viewer-
Pro 5.0.
Aoki, T., Akashi, T., Ayabe, S., 2000. Flavonoids of leguminous plants:
structure, biological activity, and biosynthesis. J. Plant Res. 113, 475–
488.
Ayabe, S., Akashi, T., Aoki, T., 2002. Cloning of cDNAs encoding P450s
in the flavonoid/isoflavonoid pathway from elicited leguminous cell
cultures. Meth. Enzymol. 357, 360–369.
Dewick, P.M., 1988. Flavonoids: advances in research since 1980. In:
Harborne, J.B. (Ed.), Isoflavonoids. Chapman and Hall, New York,
pp. 125–209.
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.
Gang, D.R., Lavid, N., Zubieta, C., Chen, F., Beuerle, T., Lewinsohn, E.,
Noel, J.P., Pichersky, E., 2002. Characterization of phenylpropene O-
methyltransferases from sweet basil: facile change of substrate spec-
ificity and convergent evolution within a plant O-methyltransferase
family. Plant Cell 14, 505–519.
Ober, D., 2005. Seeing double: gene duplication and diversification in
plant secondary metabolism. Trends Plant Sci. 10, 444–449.
4.4. Phylogenetic and evolutionary analyses
`
Perriere, G., Gouy, M., 1996. WWW-query: an on-line retrieval system for
biological sequence banks. Biochimie 78, 364–369.
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.
Amino acid sequences of the OMTs were aligned using
Clustal W (Thompson et al., 1994). The phylogenetic tree
was constructed with the neighbor-joining method (Saitou
`
and Nei, 1987) and displayed by NJPlot software (Perriere
Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for
reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.
Sawada, Y., Ayabe, S., 2005. Multiple mutagenesis of P450 isoflavonoid
synthase reveals a key active-site residue. Biochem. Biophys. Res.
Commun. 330, 907–913.
Sawada, Y., Kinoshita, K., Akashi, T., Aoki, T., Ayabe, S., 2002. Key
amino acid residues required for aryl migration catalyzed by the
cytochrome P450 2-hydroxyisoflavanone synthase. Plant J. 31, 555–
564.
and Gouy, 1996). For the reconstruction of ancestral
sequences, DNA sequences of the OMTs were aligned
according to the protein sequence alignment using Codon-
Align software (http://www.sinauer.com/hall/), and maxi-
mum likelihood analyses under codon models were
performed using the codeml program in the PAML pack-
age (Yang, 1997).
Shimada, N., Aoki, T., Sato, S., Nakamura, Y., Tabata, S., Ayabe,
S., 2003. A cluster of genes encodes the two types of chalcone
isomerase involved in the biosynthesis of general flavonoids and
legume-specific 5-deoxy(iso)flavonoids in Lotus japonicus. Plant
Physiol. 131, 941–951.
Shimada, N., Sasaki, R., Sato, S., Kaneko, T., Tabata, S., Aoki, T.,
Ayabe, S., 2005. A comprehensive analysis of six dihydroflavonol 4-
reductases encoded by a gene cluster of the Lotus japonicus genome. J.
Exp. Bot. 56, 2573–2585.
Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL
W: improving the sensitivity of progressive multiple sequence
alignment through sequence weighting, position-specific gap
penalties and weight matrix choice. Nucleic Acids Res. 22,
4673–4680.
Acknowledgements
We thank Dr. T. Aoki of Nihon University for useful
discussions. This work was funded by a Grant-in-Aid for
Scientific Research (C) (No. 15510183) from the Japan
Society for the Promotion of Science and the 21st Century
COE Program of Ministry of Education, Culture, Sports,
Science and Technology of Japan, and partially supported
by Grant No. DE-FG03-00ER15078 from the US Depart-
ment of Energy.
VanEtten, H.D., Matthews, P.S., Mercer, E.H., 1983. (+) Maackiain and
(+) medicarpin as phytoalexins in Sophora japonica L. and identifica-
tion of the (ꢀ) isomers by biotransformation. Phytochemistry 22,
2291–2295.
References
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.
Yang, Z., 1997. PAML: a program package fro phylogenetic estimation
from analysis by maximum likelihood. Comput. Appl. Biosci. 13, 555–
556.
Zubieta, C., He, X.Z., Dixon, R.A., Noel, J.P., 2001. Structures of two
natural product methyltransferases reveal the basis for substrate
specificity in plant O-methyltransferases. Nature Struct. Biol. 8, 271–
279.
Akashi, T., Aoki, T., Ayabe, S., 1999. Cloning and functional expression
of a cytochrome P450 cDNA encoding 2-hydroxyisoflavanone syn-
thase involved in biosynthesis of the isoflavonoid skeleton in licorice.
Plant Physiol. 121, 821–828.
Akashi, T., Aoki, T., Ayabe, S., 2005. Molecular and biochemical
characterization of 2-hydroxyisoflavanone dehydratase. Involvement
of carboxylesterase-like proteins in leguminous isoflavone biosynthe-
sis. Plant Physiol. 137, 882–891.
Akashi, T., Sawada, Y., Shimada, N., Sakurai, N., Aoki, T., Ayabe, S.,
2003. cDNA cloning and biochemical characterization of S-adenosyl-