Highly conserved progesterone 5β-reductase genes (P5βR) from 5β-cardenolide-free and 5β-cardenolide-producing angiosperms

Article abstract of DOI:10.1016/j.phytochem.2010.06.004

Most cardenolides used in the therapy of cardiac insufficiency are 5β-configured and thus the stereo-specific reduction of the Δ^4,5-double bond of a steroid precursor is a crucial step in their biosynthesis. This step is thought to be catalysed by progesterone 5β-reductases. We report here on the isolation of 11 progesterone 5β-reductase (P5βR) orthologues from 5β-cardenolide-free and 5β-cardenolide-producing plant species belonging to five different angiosperm orders (Brassicales, Gentianales, Lamiales, Malvales and Solanales). Amino acid sequences of the P5βR described here were highly conserved. They all contain certain motifs qualifying them as members of a class of stereo-selective enone reductases capable of reducing activated CC double bonds by a 1,4-addition mechanism. Protein modeling revealed seven conserved amino acids in the substrate-binding/catalytic site of these enzymes which are all supposed to exhibit low substrate specificity. Eight P5βR genes isolated were expressed in Escherichia coli. Recombinant enzymes reduced progesterone stereo-specifically to 5β-pregane-3,20-dione. The progesterone 5β-reductases from Digitalis canariensis and Arabidopsis thaliana reduced activated CC double bonds of molecules much smaller than progesterone. The specific role of progesterone 5β-reductases of P5βRs in cardenolide metabolism is challenged because this class of enone reductases is widespread in higher plants, and they accept a wide range of enone substrates.

Full text of DOI:10.1016/j.phytochem.2010.06.004

Phytochemistry 71 (2010) 1495–1505
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Highly conserved progesterone 5b-reductase genes (P5bR) from
5b-cardenolide-free and 5b-cardenolide-producing angiosperms
Peter Bauer a, Jennifer Munkert a, Margareta Brydziun a, Edyta Burda b, Frieder Müller-Uri a, Harald Gröger b,
Yves A. Muller c, Wolfgang Kreis a,
*
a Department of Biology, University of Erlangen-Nuremberg, Staudtstr. 5, 91058 Erlangen, Germany
b Department of Chemistry and Pharmacy, University of Erlangen-Nuremberg, Henkestr. 42, 91054 Erlangen, Germany
c Department of Biology, University of Erlangen-Nuremberg, Henkestr. 91, 91054 Erlangen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 April 2010
Received in revised form 2 June 2010
Available online 1 July 2010
Most cardenolides used in the therapy of cardiac insufficiency are 5b-configured and thus the stereo-spe-
cific reduction of the
D
4,5-double bond of a steroid precursor is a crucial step in their biosynthesis. This
step is thought to be catalysed by progesterone 5b-reductases. We report here on the isolation of 11 pro-
gesterone 5b-reductase (P5bR) orthologues from 5b-cardenolide-free and 5b-cardenolide-producing plant
species belonging to five different angiosperm orders (Brassicales, Gentianales, Lamiales, Malvales and
Solanales). Amino acid sequences of the P5bR described here were highly conserved. They all contain cer-
tain motifs qualifying them as members of a class of stereo-selective enone reductases capable of reduc-
ing activated C@C double bonds by a 1,4-addition mechanism. Protein modeling revealed seven
conserved amino acids in the substrate-binding/catalytic site of these enzymes which are all supposed
to exhibit low substrate specificity. Eight P5bR genes isolated were expressed in Escherichia coli. Recom-
binant enzymes reduced progesterone stereo-specifically to 5b-pregane-3,20-dione. The progesterone
5b-reductases from Digitalis canariensis and Arabidopsis thaliana reduced activated C@C double bonds
of molecules much smaller than progesterone. The specific role of progesterone 5b-reductases of P5bRs
in cardenolide metabolism is challenged because this class of enone reductases is widespread in higher
plants, and they accept a wide range of enone substrates.
Keywords:
Digitalis
Plantaginaceae
Progesterone 5b-reductase
Enone reductase
Substrate promiscuity
Cardenolide
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
that a metabolic grid rather than a pathway should be used to dis-
play and describe cardenolide biosynthesis (Kreis and Müller-Uri,
5b-Cardenolides are C23-steroids containing a butenolide ring at
C-17 and are derived from mevalonic acid via phytosterol and
pregnane intermediates (Fig. 1). Further structural features are
the cis–trans–cis configuration of the four annealed carbon rings
and a hydroxyl function at C-14b. Several 5b-cardenolides, such
as digoxin or digitoxin, are of special interest since they are used
in the therapy of cardiac insufficiency in humans.
The reaction converting sterols into pregnenolone is thought to
be catalysed by a mitochondrial cytochrome P450-dependent side
chain cleaving enzyme (P450scc, CYP11A in animals) (Lindemann
and Luckner, 1997) although no evidence of such a P450 has been
found in plants as yet (Ohnishi et al., 2009). Subsequently, preg-
nenolone has to be modified in several steps and condensed with
a C2 unit to yield the 5b-cardenolide genin (Fig. 1). The preferred
sequence of the individual biosynthetic steps leading to 5b-carde-
nolides is not yet clear and more than one pathway may be oper-
ative (Maier et al., 1986; Kreis et al., 1998) and it was suggested
2010).
The stereo-specific reduction of the
D
4,5-double bond of puta-
tive steroid precursors, such as progesterone, was investigated
thoroughly. A progesterone 5b-reductase supposed to be involved
in 5b-cardenolide metabolism (P5bR; EC 1.3.1.3) was isolated
(Gärtner et al., 1994) and genes encoding this enzyme were cloned
from several Digitalis species and functionally expressed in Esche-
richia coli (Herl et al., 2006a,b, 2008; Gavidia et al., 2007). The
orthologues Arabidopsis thaliana VEP1 gene encodes a protein
which seems to be required for vascular strand development (Jun
et al., 2002) but was also shown to be capable of reducing proges-
terone and other
D
4,5-steroids stereo-specifically in vitro (Herl
et al., 2009). Only recently, Perez-Bermudez et al. (2010) reported
on the isolation and functional expression of a novel progesterone
5b-reductase (termed P5bR2) from Digitalis purpurea that can be
induced by stress.
The occurrence of genes coding for progesterone 5b-reductases
in cardenolide-free species together with our observation that
some P5bRs are highly substrate-promiscuous in vitro (Herl et al.,
2006a,b; Burda et al., 2009; Schebitz et al., 2010) highlights the
* Corresponding author. Tel.: +49 91318528241; fax: +49 91318528243.
E-mail address: wkreis@biologie.uni-erlangen.de (W. Kreis).
0031-9422/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2010.06.004

1496
P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
Fig. 1. Putative pathways leading from phytosterols to 5b-cardenolides such as digitoxigenin. P5bR/P5bR2 – progesterone 5b-reductases.
existence of a significant gap in our present understanding of the
exact physiological function of P5bRs. We here isolated further
P5bR genes, demonstrate that their substrate-binding site is highly
conserved with regard to the amino acids flanking the two catalytic
amino acids and assume that substrate promiscuity is a likely
property of all P5bRs.
et al., 2009). Therefore these compounds may have a wider distri-
bution than previously assumed.
The unpredictable occurrence of 5b-cardenolides in the angio-
sperms raises the question whether this trait has evolved only once
or several times during evolution. The analysis of genes suggested
to be involved in 5b-cardenolide formation may help to address
this question.
2. Results and discussion
2.1. Occurrence of 5b-cardenolides
2.2. Occurrence of progesterone 5b-reductases
5b-Cardenolides (Fig. 1) have been found in ca. 60 genera of the
angiosperms (Singh and Rastogi, 1970; Melero et al., 2000; Kreis
and Müller-Uri, 2010). So far they have been reported to occur in
the monocots (Poales, Asparagales, Liliales), the basal eudicots
(Ranunculales), the rosids (Crossosomatales, Myrtales, Celastrales,
Malpighiales, Fabales, Rosales, Brassicales, Malvales) and the aster-
ids (Gentianales, Lamiales, Solanales, Asterales). No records were
found for the magnoliids (Fig. 2). About half of the genera investi-
gated up to today belong to the Gentianales. It appears that 5b-car-
denolides have a higher prevalence in phylogenetical younger
angiosperms. Only recently, 5b-cardenolides were reported in the
Asterales (Wang et al., 2007) and Crossosomatales (Klausmeyer
Progesterone 5b-reductases are thought to catalyse an impor-
tant step in the 5b-cardenolide biosynthesis, namely the conver-
sion of progesterone to 5b-pregnane-3,20-dione (Gärtner et al.,
1990). Therefore P5bR was regarded to catalyse the first committed
step in 5b-cardenolide biosynthesis (Gärtner and Seitz, 1993). Two
different progesterone 5b-reductases (P5bR, P5bR2) have been
shown to occur in the genus Digitalis. Following the nomenclature
of the short-chain dehydrogenase/reductase (SDR) super-family of
proteins (Persson et al., 2009) Perez-Bermudez et al. (2010) pro-
posed that the family designation for progesterone 5b-reductases
should be SDR75R and the protein designations should be
SDR75R1 and SDR75R2 for P5bR and P5bR2, respectively. We here

P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
1497
Amborealles
Nymphaeales
Austrobaileyales
Chloranthales
Ceratophyllales
Canellales
Piperales
Laurales
Magnoliids
Magnoliales
Acorales
Alismatales
Petrosaviales
Dioscoreales
Pandanales
Liliales
Monocots
Asparagales
Arecales
Poales
Commelinids
Commelinales
Zingiberales
Ranunculales
Sabiales
Proteales
Trochodendrales
Buxales
Gunnerales
Berberidopsidales
Dilleniales
Caryophyllales
Santalales
Saxifragales
Vitales
Crossosomatales
Geraniales
Myrtales
Zygophyllales
Celastrales
Malpighiales
Oxalidales
Fabales
Rosales
Cucurbitales
Fagales
Eudicots
Eurosids I
Eurosids II
Rosids
Brassicales
Malvales
Sapindales
Cornales
Ericales
Garryales
Gentianales
Lamiales
Solanales
Boraginales
Aquifoliales
Apiales
Euasterids I
Euasterids II
Asterids
Asterales
Dipsacales
Fig. 2. Distribution of 5b-cardenolides in angiosperms. Orders reported to contain 5b-cardenolides are marked with circles (d,s). Filled circles (d) and open triangles (
D)
indicate that members of this order have been demonstrated previously or during this study to contain P5bR genes.
use the established terms P5bR and P5bR2 or speak of P5bR-like
genes and P5bR-like enone reductases.
A high number of P5bR-like proteins in the public domain dat-
abases could be identified using published sequences of recombi-
nant P5bRs from Digitalis species as a template (Gavidia et al.,
2007 and own BLAST search). Bacterial P5bR-like genes form sepa-
rate clusters (not shown) and among the angiosperms two individ-
ual clusters can be identified, one containing Digitalis P5bR, the

1498
P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
100
Populus deltoides ABK96577.1
Populus trichocarpa XP 002327863.1
Vitis vinifera XP 002276159.1
Digitalis purpurea ACZ66261.1
Oryza sativa NP 001060201.1
Oryza sativa BAC20032.1
100
100
Cluster I
80
99
85
Oryza sativa EEE67538.1
58
Sorghum bicolor XP 002449263.1
Zea mays ACN25929.1
100
Sorghum bicolor XP 002465092.1
Oryza sativa NP 001050440.1
Oryza sativa EAY90583.1
100
100
100
Ricinus communis XP 002518657.1
Medicago truncatula ABD33273.2
Medicago truncatula ABD33275.2
Arabidopsis thaliana ABU55811.1
Arabidopsis thaliana CAA68126.1
Arabidopsis thaliana NP 194153.1
Arabidopsis thaliana NP 001078438.1
Erysimum rhaeticum GU354237
Erysimum crepidifolium GU354236
Populus deltoides ABK96579.1
Populus trichocarpa XP 002326975.1
Populus trichocarpa XP 002326976.1
Populus trichocarpa XP 002304533.1
Populus trichocarpa XP 002302420.1
Populus trichocarpa XP 002326977.1
Populus trichocarpa XP 002302422.1
Populus trichocarpa XP 002326973.1
Ricinus communis XP 002510123.1
Vitis vinifera CAN63254.1
61
42
87
100
100
98
98
40
100
93
Cluster II
84
99
86
43
62
100
100
100
Vitis vinifera XP 002285152.1
Isoplexis canariensis ABB36651.1
Digitalis lanata AAS76634.1
99
98
65
Plantago major GU354233
Mentha piperita GU451677
99
Nicotiana tabacum B BAH47641.1
Nicotiana tabacum A BAH47640.1
Atropa belladonna GU451678
Corchorus olitorius GU479997
Nerium orleander GU354232
100
97
47
Hoya carnosa GU354231
97
Calotropis procera GU479996
Gomphocarpus fruticosus GU354238
Asclepias curassavica GU354230
Picea sitchensis ABK24388.1
99
100
94
Outgroup
Fig. 3. Rooted neighbour-joining tree (Saitou and Nej, 1987) for P5bR-like enone reductases (strictly conserved motifs IV = NFYYxxED and V = TGxKxYxG). Data from protein
blast search and the new sequences described here. The optimal tree with the sum of branch length = 5.13436521 is shown. The percentage of replicate trees in which the
associated taxa clustered together in the bootstrap test (500 replicates) greater than 40% are shown next to the branches (Felsenstein, 1985). The evolutionary distances were
computed using the Poisson correction method (Zuckerkandl and Pauling, 1965) and are in the units of the number of amino acid substitutions per site. All positions
containing alignment gaps and missing data were eliminated only in pairwise sequence comparisons (pairwise deletion option). There were a total of 1050 positions in the
final dataset. Cladistic analyses were conducted in MEGA4.1 (Tamura et al., 2007). Triangles (
filled triangles ( ) indicate those species where the respective P5bR was functionally expressed. P5bR2 from Digitalis purpurea (ACZ66261) was included since it functionality
as a progesterone 5b-reductase was shown although the motifs IV and V are slightly variant (Perez-Bermudez et al., 2010; Supplementary Fig. 2).
N/D) indicate the species from which the new P5bR genes were cloned and the
N
other one containing Digitalis P5bR2 (Fig. 3). Although 5b-cardeno-
lides are present in many angiosperms so far no entries (except for
Digitalis) were found in the public databases indicating the pres-
ence of P5bR genes in these species. Therefore we here aimed at
the isolation of P5bR genes from a set of 5b-cardenolide-producing
species and close cardenolide-free relatives (Table 1) in order to
address the question of relevance of progesterone 5b-reductases
in 5b-cardenolide metabolism.
fruticosus, Hoya carnosa, Mentha piperita, Nerium oleander and Plan-
tago major. All new cDNA sequences reported here have been sub-
mitted to GenBank™ with the accession numbers listed in Table 1.
After sequencing of the cDNAs the corresponding protein se-
quences were deduced and aligned against the recombinant P5bR
from D. lanata (Herl et al., 2006a). The deduced proteins share
68–96% sequence identity with the D. lanata P5bR and 50–54%
identity with the D. purpurea P5bR2, respectively. A partial align-
ment presenting the SDR motifs I–VI of all proteins is shown in Ta-
ble 2. Each of the genes isolated in the present study codes for
enzymes containing all of these motifs. The full alignments (new
P5bR gene products versus D. lanata P5bR and versus P5bR2,
respectively) are provided as Supplementary Figs. 1 and 2.
Only 21 out of 85 deduced protein sequences obtained from a
homology search (e values < eÀ70 compared to the D. lanata P5bR
protein = ca. 50% identity) did not contain the conserved P5bR-spe-
cific variants of the motifs IV and V, namely NFYYxxED (containing
the catalytic tyrosine residue) and TGxKxYxG (containing the cat-
2.3. Isolation, cloning and sequence analysis of P5bR orthologues
The PCR amplification of putative P5bR genes was achieved with
a set of primers deduced from GenBank™ entries for Digitalis lanata
P5bR (Herl et al., 2006a), the A. thaliana At5b-StR (Herl et al., 2009)
and the enone reductase gene of Nicotiana tabacum (Matsushima
et al., 2008). We succeeded in the isolation of cDNAs from Asclepias
curassavica, Atropa belladonna, Calotropis procera, Corchorus olitori-
us, Erysimum crepidifolium and Erysimum rhaeticum, Gomphocarpus

P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
1499
Table 1
P5bR genes isolated here and those described previously with proof of function as progesterone 5b-reductase.
Species investigated
Order
Accession No.
ORF (nts)
5b-Cardenolides present
Functional P5bR
Arabidopsis thalianaa
Erysimum crepidifoliumb
Erysimum rhaeticumb
Asclepias curassavicab
Calotropis procerab
Gomphocarpus fruticosusb
Hoya carnosab
Brassicales
Brassicales
Brassicales
Gentianales
Gentianales
Gentianales
Gentianales
Gentianales
Lamiales
Lamiales
Lamiales
Lamiales
Lamiales
EF579963
GU354236
GU354237
GU354230
GU479996
GU354238
GU354231
GU354232
AY585867
AJ310673
GU062787
DQ218315
GU451677
GU354233
GU479997
GU451678
AB488494
1167
1170
1170
1164
1164
1167
1164
1164
1170
1167
1182
1170
1170
1199
1164
1164
1170
À
+
+
+
+
+
+
n.d.
+
+
+
+
+
+
+
+
À (5
a
-Cardenolides)
-Cardenolides)
+
À (5
À
+
+
+
+
+
À
À
+
a
Nerium oleanderb
Digitalis lanatac
Digitalis purpuread
Digitalis purpureae
Digitalis canariensisf
Mentha piperitab
Plantago majorb
Lamiales
Malvales
Solanales
Solanales
+
Corchorus olitoriusb
Atropa belladonnab
Nicotiana tabacumg
n.d.
n.d.
+
À
À
n.d. – not determined.
a
Herl et al., 2009.
This paper.
Herl et al., 2006a.
Gavidia et al., 2007.
Perez-Bermudez et al., 2010.
Herl et al., 2006b.
Matsushima et al., 2008.
b
c
d
e
f
g
Table 2
Alignment of the P5bR sequence motifs defined by Thorn et al. (2008) in the restricted short-chain dehydrogenase/reductases of the P5bR-like enone
reductases. Conserved amino acids are highlighted. The numeration is given for orientation. The complete sequence alignments are shown in Supplementary
Fig. 1.
alytic lysine residue). We also identified several P5bR-like genes by
aligning D. lanata P5bR with proteins possessing higher e values
but containing the conserved P5bR motifs IV and V.
The cladogramme generated from all sequences showed that
one cluster (Cluster II) of P5bRs correlated nicely with the assumed
phylogenetic relationship of the species in the cluster. This implied
that these genes have evolved from a common ancestral gene. A
second cluster (Cluster I) contained progesterone 5b-reductase
genes of Populus, Vitis and D. purpurea. The data set of Cluster I
genes is too small to infer phylogenetic relationship. Interestingly,
all three genera also contained paralogues in Cluster II. The occur-
rence of multiple forms of progesterone 5b-reductase genes in Pop-
ulus and Arabidopsis with paralogues on different chromosomes is
noteworthy. P5bRs (Cluster I) with a high degree of similarity
and the motifs IV and V conserved existed in all 5b-cardenolide-
containing genera investigated here (Erysimum, Calotropis, Nerium,
The P5bRs deduced here and proteins possessing all P5bR motifs
were used to investigate their relationship. It was tested whether
structural similarity correlated with a phylogenetic relationship
of the species included in this analysis. A possible evolutionary his-
tory was created using the neighbour-joining method (Saitou and
Nej, 1987). The tree was rooted using Picea sitchensis as outgroup.
This seemed appropriate because the respective P. sitchensis gene
ABK 24338.1 also coded for a P5bR-like protein containing the con-
served motifs IV (NFYYxxED) and V (TgxKxYxG) which we identi-
fied as characteristic for P5bR-like proteins (Fig. 3).

1500
P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
Corchorus) but also in plants containing 5
a- but not 5b-cardeno-
performed earlier (Herl et al., 2009) and are consistent with the
lides (Gomphocarpus, Asclepias) or even in cardenolide-free species
(Hoya, Mentha, Plantago, Atropa).
expected reaction mechanism (Thorn et al., 2008). For each com-
plex, a total of 16 amino acids located within 4.5 Å of either any
progesterone atoms or atoms from the ribose and nicotine amide
moiety of the co-substrate NADPH were considered as participat-
ing in the substrate-binding site and thereby contributing to sub-
strate specificity (Table 2). As in the previously characterised
P5bR-like proteins conserved Y179 and K147 residues were located
in the active site. Interestingly, we found another five conserved
amino acids within this pocket, namely W106, G145, F153, M215
and F353 (Fig. 4). G145, K147 and Y179 are amino acids displayed
from the characteristic motifs IV and V (Table 3). On the other
hand, W106, F153, M215 and F353 do not appear in any of the
known sequence motifs but at the same time have a pronounced
effect on the shape of the binding site. The conserved amino acids
cluster around those parts of NADPH and progesterone that partic-
ipate in the stereo-selective C@C double bond reduction in this
way forming the conserved apex of an otherwise variable cone.
No conserved amino acids could be found in the more remote parts
of the catalytic pocket. To summarise, the enzymes under investi-
gation not only are highly conserved with regard to certain motifs
but also with regard to several individual amino acids in the
catalytic/substrate-binding site.
When looking at the structure of the catalytic/substrate-binding
site numerous indications can be found that progesterone is not
the unique substrate for P5bRs in vivo, as was already deduced
from the enzyme activity data (Herl et al., 2006a; Burda et al.,
2009; Schebitz et al., 2010). For example, when docking progester-
one into the substrate-binding site it becomes apparent that the
bottom of this site is quite hydrophilic. This polarity is not
mirrored in progesterone. In some of the P5bRs asparagine-205 is
replaced by methionine and indeed a N205M mutant of a recombi-
nant D. lanata P5bR shows an increased progesterone reductase
activity when compared to wild type (data not shown). On the
other hand, an asparagine residue is found at this position in many
cardenolide-producing species. Hence, these P5bRs seem not to be
optimised for the conversion of progesterone. These structural
findings lead to the hypothesis that Cluster II P5bRs are
substrate-promiscuous enone reductases. We suggest that these
enzymes may be regarded as ‘‘pontoon” elements which can be
utilised in various biochemical pathways or may contribute to
Progesterone 5b-reductase genes coding for enzymes showing
the conserved P5bR motifs IV and V, were up to now reported for
11 genera of angiosperms only. We here extended the list of genera
shown to possess P5bR genes by another 10 genera. Most of the
P5bRs isolated so far were from plant orders containing species
capable of producing 5b-cardenolides (Fig. 3). However, the occur-
rence of P5bR genes seems not to be correlated with the occurrence
of 5b-cardenolides. Especially the occurrence of P5bR homologues
in gymnosperms is exciting since cardiac glycosides have so far
not been reported from this class of seed plants. Although further
experiments are required, our results and those of others (Perez-
Bermudez et al., 2010) indicate that P5bR Cluster II orthologues
may be present in many plant species.
2.4. Molecular modeling
Progesterone 5b-reductases are members of the large SDR
super-family of enzymes (Jörnvall et al., 1995). All P5bR genes iso-
lated here shared the characteristic features describing the new
class of SDRs characterised by Thorn et al. (2008) on the basis of
X-ray data. They can be classified as stereo-specific enone reduc-
tases capable of reducing activated C@C double bonds by a 1,4-
addition mechanism. A hydride is transferred from NADPH to a
carbonyl-activated C@C double bond and the carbonyl then pro-
tonated. In this way an enol intermediate is formed which tautom-
erises to the final product. Gavidia et al. (2007) erroneously
suggested that the second tyrosine residue from the NFYYxxED
motif acts as the catalytic base but only the experimentally deter-
mined crystal structure combined with site-directed mutagenesis
(Thorn et al., 2008) revealed that it is the first tyrosine that is cru-
cial for catalysis. Experiments in which the catalytic tyrosine resi-
due was replaced by alanine or phenylalanine completely
inactivated the protein (Thorn et al., 2008) whereas other muta-
tions in the NFYYxxED motif did not (own unpublished results).
We here modeled the active sites and amino acid residues of the
substrate-binding site using D. lanata P5bR (2V6G) as well as the
progesterone-docked D. lanata structure as templates. The result-
ing complexes agree nicely with the docking experiments
Fig. 4. (A) Catalytic pocket of P5bR-like enone reductases modeled on the crystal structure 2V6G (D. lanata P5bR). All amino acids located within 4.5 Å of either any
progesterone atoms or atoms from the ribose and nicotine amide moiety of the cofactor NADPH were considered as participating in the substrate-binding site. In addition to
the catalytic Y179 and K147 residues another five conserved amino acids could be identified within this pocket, namely W106, G145, F153, M215 and F353. The amino acids
suggested to be part of the binding pocket are shown in Table 2. B. The frequency of occurrence of a certain amino acid in the variable parts of the pocket is symbolised by the
font size used. Colours: conserved amino acids are shown (A) or highlighted (B) yellow. Green, red, blue, orange and grey lines indicate the different parts of the protein
backbone contributing to the substrate-binding pocket. Substrate (progesterone) and co-substrate (NADPH) are shown in magenta.

P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
1501
Table 3
Amino acids suggested to be part of the binding pocket of P5bR-like enone reductases. All of the protein
sequences deduced here are shown together with those of the P5bRs of A. thaliana (Herl et al., 2009), D.
lanata (Herl et al., 2006a) and D. canariensis (Herl et al., 2006b). Structurally conserved amino acids are
highlighted.
occult biosynthetic capacities recently defined as ‘‘silent metabo-
lism” (Lewinsohn and Gijzen, 2009).
molecules bearing an activated C@C double bond such as monocy-
clic enones and acyclic enoate esters. The essential structural ele-
ment for substrates of P5bRs seems to be an activated C@C
double bond. This structural element is present in several natural
products, such as isopiperitenone and pulegone (menthol biosyn-
thesis), which all have a biosynthetic counterpart with the C@C
double bond eliminated. We therefore suggest that the P5bRs de-
scribed so far should be designated ‘‘enone reductases” like the
one isolated from tobacco cell cultures (Matsushima et al., 2008).
This particular enone reductase contains all structural require-
ments to qualify it as a member of the P5bR-like SDRs of the
P5bR Cluster II.
The substrate preferences of two of the recombinant P5bRs
(those from A. thaliana and Digitalis canariensis) were investigated
and compared. Both enzymes accepted progesterone and a broad
variety of small cyclic and acyclic substrates bearing an activated
C@C double bond (Table 4). The recombinant A. thaliana P5bR re-
duced all substrates ca. 10 times faster than the D. canariensis en-
zyme. This corroborates previous findings using steroid substrates
only (Herl et al., 2009). Otherwise, the substrate preferences of the
two P5bRs tested here were very similar. With both enzymes very
high activities were measured for 2-cyclohexenone which repre-
sents a monocyclic substructure of progesterone. But-1-en-3-one,
which may be regarded as the smallest possible mimic of proges-
terone, was reduced very efficiently, exceeding the relative activity
measured for progesterone by a factor of 28. The high enantio-
selectivity of the reaction, as demonstrated here for progesterone,
could also be seen with isophorone (ee > 99%), a very small prochi-
ral progesterone mimic (Burda et al., 2009). It is assumed that all
P5bRs of the Cluster II are substrate-promiscuous enzymes and
that progesterone cannot be considered a unique substrate.
Kinetic constants of the new enzymes described here have not
been collected as yet but have already been published for the
P5bR-like enone reductases of, e.g., D. lanata, D. purpurea, D. canari-
ensis and A. thaliana (Herl et al., 2006a,b, 2009; Gavidia et al.,
2007). We are aware of the fact that even the detailed biochemical
2.5. Heterologous expression of P5bR genes
The conditions described previously for a recombinant form of
D. lanata P5bR (Herl et al., 2006a) were applied and optimised for
expression of the new P5bR genes. The recombinant enzymes were
Ni–NTA-purified and analysed for functionality applying the meth-
ods described earlier (Herl et al., 2006a,b, 2008, 2009). Showing
data for rMpP5bR, a recombinant form of M. piperita P5bR, Fig. 5
exemplifies the data sets collected for each expression experiment,
including SDS PAGE, GC–MS and TLC analysis. TLC was used as a
simple and fast screening tool whereas GC–MS analysis proved
that only 5b-configured pregnanes were formed (Fig. 5).
Eight P5bR orthologues isolated from cardenolide-free plants as
well as from cardenolide-containing (either 5b- or 5a-cardeno-
lides) plants were successfully expressed in E. coli and demon-
strated to reduce progesterone enantio-specifically to 5b-
pregnane-3,20-dione. Some of the P5bR proteins obtained from
cardenolide-free plants proved to be even more active towards
progesterone than the respective proteins from cardenolide-con-
taining plants (data not shown). P5bR homologues were expected
to occur in 5b-cardenolide-free plants since the biochemical char-
acterisation of the VEP1 gene product of A. thaliana already indi-
cated that P5bR enzymes from non-cardenolide plants can be
more efficient than the Digitalis P5bR-like enone reductases inves-
tigated so far (Herl et al., 2009). The activity data presented here
confirm these findings (Table 4).
When investigating substrate preferences and kinetic proper-
ties of rDlP5bR, the recombinant, His-tagged form of a D. lanata
P5bR, it was found that not only progesterone but also testoster-
one, 4-androstene-3,17-dione, cortisol, cortisone, canarigenone
and 23-nor-4,20(22)E-choladienic acid-3-one (Herl et al., 2006a;
Schebitz et al., 2010) were accepted as substrates. It has been dem-
onstrated that the P5bR of A. thaliana is capable of reducing small

1502
P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
Fig. 5. Routine analysis exemplified with recombinant Mentha piperita P5bR (rMpP5bR). (A) TLC analysis of P5bR activity: 1 – reference progesterone (1; Rf = 0.5), 4 –
reference 5b-pregnane-3,20-dione (2; Rf = 0.6);+ – enzyme assay with the Ni–NTA-purified rMpP5bR after 2 h; À – heat-inactivated control. (B) Expression of rMpP5bR in
E. coli analysed on SDS PAGE (12%). M – molecular weight marker. P – purified rMpP5bR after Ni–NTA affinity chromatography. An arrow marks the recombinant protein at
ꢀ43 kDa. (C) GC–MS analysis of a standard mixture of 1 – 5b-pregnane-3,20-dione, Rt = 23.70 min; 2 – isoprogesterone, Rt = 24.19 min; 3–5
a-pregnane-3,20-dione,
Rt = 24.55 min; 4–progesterone, Rt = 25.93 min. (D) P5bR enzyme assay (2 h) – Enantio-selective reduction of progesterone using Ni–NTA-purified rMpP5bR.
Table 4
in more than just 5b-cardenolide biosynthesis. One may even chal-
Relative and specific activities of P5bRs from D. canariensis and A. thaliana towards
lenge the role of P5bR in the cardenolide pathway(s) and the exact
physiological role of P5bR still remains to be elucidated. Since sub-
strate preferences of P5bR2s (Perez-Bermudez et al., 2010) have
not been reported so far, only weak arguments for a more specific
role of this particular enzyme in the formation of 5b-cardenolides
have been provided as yet. Future knock-down or knock-out stud-
ies may help to solve this issue. Analysing P5bR genes of cardeno-
lide-free mutants of cardenolide-producing species may also help
to address this question. As far as the potential biotechnological
application of enantio-selective enone reductases is concerned
the structural data collected here will help to design P5bRs with
improved or variant catalytic properties. The biochemical charac-
terisation of some of the P5bRs is in progress.
various substrates.
A. thaliana
Enzyme activity
D. canariensis
Enzyme activity
Relative Specific
(%)
Relative (%) Specific
l
kat/kg
lkat/kg
protein
protein
Progesterone
2-Cyclohexenone
3-Methyl-2-
cyclohexenone
But-1-en-3-one
100
1298
19
620
8048
118
100
1204
52
50.1
602.0
26.1
2783
12
17255
74
Not
determined
86
3,5,5-Trimethyl-2-
cyclohexenone
43.1
Ethyl acrylate
Ethylmethacrylate
75
64
467
397
61
64
30.6
32.1
3. Experimental
characterisation of a given enzyme is not sufficient for deducing its
metabolic role. However, the occurrence of closely related func-
tional P5bRs in 5b-cardenolide-free as well as in 5b-cardenolide-
containing plants lets us presume that Cluster II P5bRs are involved
3.1. Plant material, chemicals, recombinant enzymes
Seeds, seedlings and plants were provided by the Botanic Gar-
den of the University of Erlangen and identified by Walter Welß.

P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
1503
Specimens were deposited in the Erlangen University Herbarium.
Plants were grown in the greenhouse until they produced
sufficient leaf material for RNA extraction. They were watered every
other day and supplemented once a week with nutrient solution
(WUXALÒ Universaldünger, Wilhelm Haug, Ammerbuch, Germany).
Steroids were purchased from Steraloids (Newport, USA). Iso-
progesterone was synthesised according to Djerassi et al. (1956).
All other chemicals were obtained from Sigma (Munich, Germany)
or Merck (Darmstadt, Germany).
fragment into a 1.5 ml Eppendorf tube and precipitated with 75%
(v/v) isopropyl alcohol at 20 °C for 30 min. After centrifugation
(12,000g, 15 min), the DNA pellet was washed with 70% ice-cold
ethanol, again centrifuged and finally dried and re-dissolved in
sterile double-distilled water.
The cloning of the purified DNA fragment into the pQE30UA vec-
tor system was performed following the instructions of the manu-
facturer (QIAexpressionist™, Hilden, Germany). A vector to insert
ratio of 1:10 was chosen in the subsequent ligation step. Ligation
was achieved over night at 16 °C. Plasmids were transferred to
chemical competent E. coli M15 [pREP4] cells. Transformed cell col-
onies were selected on plates containing ampicillin and kanamycin.
The colonies were screened by colony PCR (Alshahni et al.,
2009) using insert specific primers and the following PCR pro-
gramme: Initial denaturation (95 °C/10 min), denaturation (95 °C/
30 s), annealing (50 °C/1 min), elongation (72 °C/2 min) for 30 cy-
cles and a final elongation step of 10 min. Plasmid DNA was iso-
lated form positive clones using the peqGold Plasmid Miniprep
Kit (Peqlab GmbH, Erlangen, Germany) and then sequenced by
GATC Biotech AG (Konstanz, Germany).
Recombinant forms of the P5bRs from A. thaliana and D. (Isopl-
exis) canariensis were prepared as described by Herl et al. (2006a,b)
and Burda et al. (2009).
3.2. RNA extraction, cDNA synthesis and PCR
Only fresh, young leaves were used for RNA isolation. The leaves
were frozen with liquid nitrogen and ground to a powder using
mortar and pestle. Complete RNA isolation was carried out using
a innuPREP Plant RNA Kit (Analytik Jena AG, Jena, Germany) fol-
lowing the manufacturer’s instructions. We continued to use the
PL buffer provided in the kit, because it yielded larger amounts
of higher quality RNA. This fact was proven by photometric analy-
sis with a NANODROP ND-1000 spectrophotometer.
3.4. Subcloning and sequencing of PCR products using the GatewayÒ
Recombination Cloning system
The synthesis of P5bR cDNA was performed employing Super-
Script™ III First-Strand Synthesis for reverse transcription PCR
(RT-PCR) Kit (Invitrogen, Karlsruhe, Germany). RT-PCR was carried
out in a Personal Cycler 20 (Biometra GmbH, Göttingen, Germany)
according to the manufacturer’s recommendation. Each reaction
In one set of experiments PCR products were cloned into an
entry vector for the GatewayÒ recombination system following
the manufacturer’s instruction (Invitrogen, Karlsruhe, Germany).
pENTR/D-TOPO, the directional TOPO entry vector was chosen for
creating the entry clone. The sequences of the primers used for
the RT-PCR contained the sites essential for the respective cloning
procedure (Gengatediv: CTA AGG AAC AAT CTT GTA ACC TT and
Gengaterer: CAA CAT GAG YTG GTG GTG GGC T). Destination
cloning into the pDEST 17 vector was carried out following the
manufacturer’s instructions (Invitrogen, Karlsruhe, Germany). All
sequences obtained were analysed for correct orientation and
reading frame. The appropriate constructs were transferred to
(50 ll total volume) contained 2.5 units peqGold Taq DNA Poly-
merase (Peqlab GmbH, Erlangen, Germany), 1Â reaction buffer S,
0.5 mM of each dNTP, 2 mM of appropriate primers and 0.2
the respective cDNA.
lg of
For standard PCR the following amplification programme was
used: an initial denaturation at 94 °C for 5 min was followed by
30 cycles of a 20 s denaturation period at 94 °C, 1 min annealing
at 50 °C and 2 min extension at 72 °C. Finally a 10 min extension
at 72 °C was added to complete the amplification. PCR products
were analysed by 1% agarose gel electrophoresis in a TAE buffer
system. Gels were stained with ethidium bromide and visualised
by illumination at UV365nm. For size determination SmartLadder
(Eurogentec GmbH, Köln, Germany) was used, producing a pattern
of 14 regularly spaced bands ranging from 200 to 10.000 bp.
Degenerated as well as standard primers were deduced from
the GenBank™ entries of the known P5bR-like genes from A. thali-
ana (AY062451.1), D. lanata (AAS93804.1) and N. tabacum
(BAH47640.1). Different combinations of primers were tested in
order to gain a specific PCR product for each gene isolation/cloning
experiment. Finally, the full-length cDNA clones were amplified
using the following primers: Soldir (ATG AGC TGG TGG GCT GG)
and Solrev (CTA AGG AAC AAC TTT GTA GGC) for A. belladonna; Ery-
dir (ATG AGT TGG TGG GGG) and Erysalrev (TAT AGT CGA CTG GAA
ATC AAG GCA CGA TCT) for E. crepidifolium and E. rhaeticum; Gendir
(ATG AGY TGG TGG TGG GCT) and Genrev (AGG AAC AAT CTT GTA
AGC CTT) for C. procera, G. fruticosus, C. olitorius, M. piperita;
Gengatedir (CAAC ATG AGY TGG TGG TGG GCT) and Gengaterev
(CTA AGG AAC AAT CTT GTA AGC CTT) for P. major, H. carnosa,
A. curassavica, N. oleander.
chemically competent E. coli DH5a cells for maintenance, amplifi-
cation and later on for expression in E. coli BL21(DE3).
3.5. Over-expression of P5bR cDNA cloned in pQE30UA or DESTINY
vector in E. coli
Bacteria were incubated at 37 °C to an OD600 of 0.5–0.7. Isopro-
pyl-b-D-thiogalactoside (IPTG) was added to induce expression of
the recombinant His-tagged P5bR. Cultivation at 37 °C for up to
6 h (1 mM IPTG) as well as cultivation at 4 °C for up to 120 h
(0.1 mM IPTG) was tested, before the His-tagged P5bR was isolated
in its native form via batch purification using Ni–NTA (nickel–
nitrilotriacetic acid) agarose following the procedure described in
the manufacturer’s instruction (QIAexpressionist™, Hilden, Ger-
many). The Ni–NTA matrix was washed extensively with 20 mM
imidazole buffer before affinity-bound proteins were eluted with
250 mM imidazole buffer. After purification of the recombinant
P5bR the reaction buffer containing 100 mM HEPES-KOH (pH
8.0), 250 mM sucrose, 2 mM EDTA and 10 mM b-mercaptoethanol
in PD-10 columns (GE Healthcare, Munich, Germany) replaced the
elution buffer. Protein was quantified according to Bradford
(1976). Protein analysis on SDS PAGE was performed as reported
earlier (Müller-Uri and Reva, 2000).
3.3. Subcloning and sequencing of PCR products using the pQE30UA
vector system
3.6. Measuring progesterone 5b-reductase activity
For subcloning RT-PCR products were purified by 1% agarose gel
electrophoresis using the ‘‘freeze and squeeze” technique (Thuring
et al., 1975). The gel slice containing the desired DNA fragment was
cut out under UV365nm after straining with ethidium bromide and
frozen at À80 °C for 5 min. The DNA was squeezed out of this gel
To demonstrate progesterone 5b-reductase activity the method
described by Stuhlemmer and Kreis (1996) and modified by Herl
et al. (2006a) was used. In a final volume of 1000
ll the standard
enzyme assay contained: 945 l (200 g/ml) purified protein,
l
l

1504
P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
6.4 mM NADP+, 32.1 mM glucose 6-phosphate, 42 nkat glucose-6-
phosphate dehydrogenase and either 0.3 mM (for TLC analysis) or
0.5 mM (for GC–MS analysis) progesterone. Heat-inactivated sam-
ples (10 min, 100 °C) served as controls. The mixture was incu-
bated in 2 ml Eppendorf tubes at 40 °C for 2 h prior to extraction
Bradford, M.M., 1976.
A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding.
Anal. Biochem. 72, 248–254.
Burda, E., Krausser, M., Fischer, G., Hummel, W., Müller-Uri, F., Kreis, W., Gröger, H.,
2009. Recombinant
D4,5-steroid 5b-reductases as biocatalysts for the reduction
of activated C@C-double bonds in monocyclic and acyclic molecules. Adv.
Synth. Catal. 351, 2787–2790.
with 1000
ll dichloromethane. The organic phase was evaporated,
Djerassi, C., Engle, R.R., Bowers, A., 1956. The direct conversion of steroidal
alcohols to 5- and 4-3-ketones. J. Org. Chem. 21, 1547–1549.
D5-3b-
D
D
the residue was dissolved in 50
ll methanol (for TLC) or 100 ll
Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the
bootstrap. Evolution 39, 783–791.
Gärtner, D.E., Seitz, H.U., 1993. Enzyme activities in cardenolide-accumulating,
mixotrophic shoot cultures of Digitalis purpurea L.. J. Plant Physiol. 141, 269–
275.
dichloromethane (for GC–MS) and then analysed. Activity tests
were carried out three times or more.
Progesterone 5b-reductase activity was determined semiquan-
titatively using TLC which was carried out as described by Herl
et al. (2006a). For product identification samples were analysed
by GC Hewlett–Packard HP 6890 MSD Type 5890A in IE mode
using helium as the carrier gas (constant flow 1 ml/min). A fused
capillary column HP Optima five coated with 5% phenyl/95% di-
methyl polysiloxane (30 m  25 mm  0.25 mm) was used. The
GC column temperature was programmed from 200 °C (initial
equilibrium time 4 min) to 300 °C at a ramp of 4 °C/min and then
maintained at 300 °C for 6 min. The MSD was operated in full-scan
mode from m/z = 50–600. The inlet and MS transfer line tempera-
Gärtner, D.E., Wendroth, S., Seitz, H.U., 1990. A stereo-specific enzyme of the
putative biosynthetic pathway of cardenolides. Characterization of
progesterone 5b-reductase from leaves of Digitalis purpurea L.. FEBS Lett. 271,
239–242.
a
Gärtner, D.E., Keilholz, W., Seitz, H.U., 1994. Purification, characterization and
partial peptide microsequencing of progesterone 5b-reductase from shoot
cultures of Digitalis purpurea. Eur. J. Biochem. 225, 1125–1132.
Gavidia, I., Tarrio, R., Rodriguez-Trelles, F., Perez-Bermudez, P., Seitz, H.U., 2007.
Plant progesterone 5b-reductase is not homologous to the animal enzyme.
Molecular evolutionary characterization of P5bR from Digitalis purpurea.
Phytochemistry 68, 853–864.
Herl, V., Fischer, G., Müller-Uri, F., Kreis, W., 2006a. Molecular cloning and
heterologous expression of progesterone 5b-reductase (P5bR) from Digitalis
lanata EHRH. Phytochemistry 67, 225–231.
tures were maintained at 300 °C. Sample injection (2 ll) was in
splitless mode. Pregnanes were identified by comparing their Rt
and fragmentation patterns with those of authentic compounds.
Spectrophotometric measurement of progesterone 5b-reduc-
tase activity was carried out as described by Burda et al. (2009).
In contrast to the general procedure described above no NADPH-
regenerating was added to the enzyme assay. The conversion of
NADPH to NADP+ was followed at 340 nm in the presence of the
respective substrates (10 mM) and the co-substrate (12.5 mM).
Relative activities towards the various substrates were determined
by comparing the amount of NADP+ released from NADPH with the
activity measured for progesterone which was set to equal 100%.
Herl, V., Fischer, G., Bötsch, R., Müller-Uri, F., Kreis, W., 2006b. Molecular cloning
and expression of progesterone 5b-reductase (P5bR) from Isoplexis canariensis.
Planta Med. 72, 1163–1165.
Herl, V., Albach, D.C., Müller-Uri, F., Bräuchler, C., Heubl, G., Kreis, W., 2008. Using
progesterone 5b-reductase, a gene encoding a key enzyme in the cardenolide
biosynthesis, to infer the phylogeny of the genus Digitalis. Plant Syst. Evol. 271,
65–78.
Herl, V., Fischer, G., Reva, V.A., Stiebritz, M., Muller, Y.A., Müller-Uri, F., Kreis, W.,
2009. The VEP1 gene (At4g24220) encodes
a short-chain dehydrogenase/
reductase with 3-oxo- 4,5-steroid 5b-reductase activity in Arabidopsis thaliana
D
L.. Biochimie 91, 517–525.
Jörnvall, H., Persson, B., Krook, M., Atrien, S., Gonzalez-Duarte, R., Jeffry, J., Ghosh, D.,
1995. Short-chain dehydrogenase/reductase (SDR). Biochemistry 34, 6003–
6013.
Jun, J.H., Ha, C.M., Nam, H.G., 2002. Involvement of the VEP1 gene in vascular strand
development in Arabidopsis thaliana. Plant Cell Physiol. 43, 323–330.
Klausmeyer, P., Zhou, Q., Scudiero, D.A., Uranchimeg, B., Melillo, G., Cardellina II, J.H.,
3.7. In silico analysis
Shoemaker, R.H., Chang, C., McCloud, T.G., 2009. Cytotoxic and HIV-1
a
inhibitory compounds from Crossosoma bigelovii. J. Nat. Prod. 72, 805–812.
After sequencing the PCR products (GATC AG, Konstanz, Ger-
many), the data were analysed using appropriate software pack-
ages (European Bioinformatics Institute). BLAST search of the
GenBank™ database (<http://blast.ncbi.nlm.nih.gov/BLAST.cgi>)
was used for searching and sequence analysis. The nucleotide se-
quences were translated into their corresponding amino acid se-
quence employing <http://www.expasy.ch/tools/dna.html>. Both
of them were aligned using ClustalW (<http://www.ebi.ac.uk/clu-
stalw/index.html>). Cladogrammes were constructed with the
MEGA4.1 software package (Tamura et al., 2007). SWISS-MODEL
(<http://swissmodel.expasy.org>) was used for protein structure
homology modeling. The three-dimensional structural drawings
were created with PyMOL (<http://www.pymol.org>).
Kreis, W., Müller-Uri, F., 2010. Biochemistry of sterols, cardiac glycosides,
brassinosteroids, phytoecdysteroids and steroid saponins, second ed.. In:
Wink, M. (Ed.), Biochemistry of Plant Secondary Metabolism, vol. 40 CRC
Press, Sheffield, pp. 304–363.
Kreis, W., Hensel, A., Stuhlemmer, U., 1998. Cardenolide biosynthesis in foxglove.
Planta Med. 64, 491–499.
Lewinsohn, E., Gijzen, M., 2009. Phytochemical diversity: the sounds of silent
metabolism. Plant Sci. 178, 161–169.
Lindemann, P., Luckner, M., 1997. Biosynthesis of pregnane derivatives in somatic
embryos of Digitalis lanata. Phytochemistry 46, 507–513.
Maier, M.S., Seldes, A.M., Gros, E.G., 1986. Biosynthesis of the butenolide ring of
cardenolides in Digitalis purpurea. Phytochemistry 25, 1327–1329.
Matsushima, A., Sato, Y., Otsuka, M., Watanabe, T., Yamamoto, H., Hirata, T., 2008.
An enone reductase from Nicotiana tabacum: cDNA cloning, expression in
Escherichia coli, and reduction of enones with the recombinant proteins. Bioorg.
Chem. 36, 23–28.
Melero, C.P., Medarde, M., San Feliciano, A., 2000. A short review on cardiotonic
steroids and their aminoguanidine analogues. Molecules 5, 51–81.
Müller-Uri, F., Reva, V.A., 2000. Overexpression and catalytic function of cyclophilin
18 from Digitalis lanata Ehrh. Pharm. Pharmacol. Lett. 10, 5–7.
Ohnishi, T., Yokota, T., Mizutani, M., 2009. Insights into the function and
evolution of P450s in plant steroid metabolism. Phytochemistry 70, 1918–
1929.
Acknowledgements
We thank Gabriele Fischer for excellent technical assistance,
Nadine Meitinger for a sample of isoprogesterone and Barbara
White for linguistic advice.
Perez-Bermudez, P., Moya Garcia, A.A., Tunon, I., Gavidia, I., 2010. Digitalis purpurea
P5bR2, encoding steroid 5b-reductase, is a novel defense-related gene involved
in cardenolide biosynthesis. New Phytol. 185, 687–700.
Persson, B., Kallberg, Y., Bray, J.E., Bruford, E., Dellaporta, S.L., Favia, A.D., Duarte,
R.G., Jörnvall, H., Kavanagh, K.L., Kedishvili, N., Kisiela, M., Maser, E., Mindnich,
R., Orchard, S., Penning, T.M., Thornton, J.M., Adamski, J., Oppermann, U., 2009.
The SDR (short-chain dehydrogenase/reductase and related enzymes)
nomenclature initiative. Chem.-Biol. Interact. 178, 94–98.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.phytochem.2010.06.004.
Saitou, N., Nej, M., 1987. The neighbour-joining method:
a new method for
reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.
Schebitz, P., Nothdurft, L., Hensel, A., Müller-Uri, F., Kreis, W., 2010. Norcholanic
acids as substrates for recombinant 3b-hydroxysteroid dehydrogenase and
progesterone 5b-reductase, enzymes of the 5b-cardenolide biosynthesis.
Tetrahedron Lett. 51, 367–370.
References
Alshahni, M.M., Makimura, K., Yamada, T., Satoh, K., Ishihara, Y., Takatori, K.,
Sawada, T., 2009. Direct colony PCR of several medically important fungi using
Ampdirect(R) Plus. Jap. J. Infect. Disord. 62, 164–167.
Singh, B., Rastogi, R.P., 1970. Cardenolides–glycosides and genins. Phytochemistry 9,
315–331.

P. Bauer et al. / Phytochemistry 71 (2010) 1495–1505
1505
Stuhlemmer, U., Kreis, W., 1996. Cardenolide formation and activity of pregnane-
modifying enzymes in cell suspension cultures, shoot cultures and leaves of
Digitalis lanata. Plant Physiol. Biochem. 34, 85–91.
Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary
genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol 24, 1596–1599.
Thorn, A., Egerer-Sieber, C., Jäger, C.M., Herl, V., Müller-Uri, F., Kreis, W., Muller, Y.,
2008. The crystal structure of progesterone 5b-reductase from Digitalis lanata
defines a novel class of short-chain dehydrogenases/reductases. J. Biol. Chem.
283, 17260–17269.
Thuring, R.W., Sanders, J.P., Borst, P., 1975. A freeze-squeeze method for recovering
long DNA from agarose gels. Anal. Biochem. 66, 213–220.
Wang, T.M., Hojo, T., Ran, F.X., Wang, R.F., Wang, R.Q., Chen, H.B., Cui, J.R., Shang,
M.Y., Cai, S.Q., 2007. Cardenolides from Saussurea stella with cytotoxicity
towards cancer cells. J. Nat. Prod. 70, 1429–1433.
Zuckerkandl, E., Pauling, L., 1965. Evolutionary divergence and convergence in
proteins. In: Bryson, V., Vogel, H.J. (Eds.), Evolving Genes and Proteins. Academic
Press, New York, pp. 97–166.