Discovery of a potent and selective inhibitor for human carbonyl reductase 1 from propionate scanning applied to the macrolide zearalenone

Article abstract of DOI:10.1016/j.bmc.2008.11.076

In order to extend the chemical diversity available for organic polyketide synthesis, the concept of propionate scanning was developed. We observed that naturally occurring polyketides frequently comprise not only acetate, but also some propionate as buil

Full text of DOI:10.1016/j.bmc.2008.11.076

Bioorganic & Medicinal Chemistry 17 (2009) 530–536
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Discovery of a potent and selective inhibitor for human carbonyl reductase 1
from propionate scanning applied to the macrolide zearalenone
Tobias J. Zimmermann a, Frank H. Niesen b, Ewa S. Pilka b, Stefan Knapp b,d
,
a,
Udo Oppermann b,c, , Martin E. Maier
*
*
a Institut für Organische Chemie, Universität Tübingen, Auf der Morgenstelle 18, 72076 Tübingen, Germany
b Structural Genomics Consortium, University of Oxford, ORCRB, Roosevelt Drive, Oxford, OX3 7DQ, United Kingdom
c Botnar Research Centre, Oxford Biomedical Research Unit, Oxford, OX3 7LD, United Kingdom
d Department of Clinical Pharmacology, University of Oxford, ORCRB, Roosevelt Drive, Oxford, OX3 7DQ, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to extend the chemical diversity available for organic polyketide synthesis, the concept of pro-
pionate scanning was developed. We observed that naturally occurring polyketides frequently comprise
not only acetate, but also some propionate as building blocks. Therefore our approach consists of a sys-
tematic replacement of some of the acetate building blocks during synthesis by propionate moieties,
resulting in additional methyl groups that may give rise to different properties of the polyketides. Here
we present the results of a first ‘proof of concept’ study where a novel zearalenone analogue 5 was pre-
pared that comprises an additional methyl group at C50. Key steps in the synthesis of 5 include a
Marshall–Tamaru reaction, a Suzuki cross-coupling reaction, and a Mitsunobu lactonization. Compared
to the parent zearalenone (1), analogue 5 showed reduced binding to a panel of human protein kinases
and no binding to human Hsp90. On the other hand, however, 5 turned out to be a potent (IC50 = 210 nM)
inhibitor of human carbonyl reductase 1 (CBR1).
Received 20 August 2008
Revised 26 November 2008
Accepted 29 November 2008
Available online 6 December 2008
Keywords:
Macrolides
Zearalenone
Natural product analogues
Polyketides
Carbonyl reductase 1
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
For any given natural product–receptor pair the question arises
which functional group(s) and/or substituent(s) represent key ele-
Many therapeutically useful compounds are based on natural
products,1 and natural products are often considered validated
starting points for drug design. It has been argued that a relation-
ship may exist between the enzymes that catalyze the biosynthesis
of natural products and the targets of these compounds:2 enzymes
binding a compound during its biosynthesis may comprise similar
protein fold topologies, that is, a similar arrangement of secondary
structure elements around the active site, as the proteins targeted
by the drug. It is obvious that natural products ‘encode’ the struc-
tural properties required for binding to their target proteins. There-
fore, the concept of biology-oriented synthesis (BIOS) was
developed that employs core structures delineated from natural
products as scaffolds of compound collections.3 For example,
Waldmann and co-workers were successful in the development
of biologically active molecules using BIOS.4 In addition, many
other groups have reported on the synthesis of libraries from nat-
ural product-like scaffolds.5
ments for the binding affinity and selectivity. To elucidate this,
scanning methods can provide valuable information. A common
example is alanine scanning, where the replacement of individual
amino acids within a peptide by L-alanine often reveals the relative
importance of these side chain groups.6,7 A fluorine scan applied to
an inhibitor of thrombin showed the FꢀꢀꢀC@O interaction to be of
special importance for the efficacy.8 In order to map a binding site,
affinity labelling and related approaches have been used.9
In natural products, such as terpenes or polyketides methyl
groups represent common substituents in addition to hydroxyl- or
keto groups. A somewhat surprising finding with many polyke-
tide-based macrolactones is the exclusive use of acetate as building
block. A prototypical example is zearalenone (1) (Fig. 1) which be-
longs to the wide-spread resorcylic lactones (RAL).10 The biosynthe-
sis of benzolactones of the zearalenone type is based on the Claisen-
condensation of nine acetates.11 The intermediate A undergoes an
intramolecular aldol condensation to form the benzoic acid B.
Other macrolactones such as dictyostatin (2), on the other hand,
are made from acetate and propionate (Fig. 2), whilst the macrolac-
tone part of erythromycin, erythronolide B (3), is entirely made
from propionate entities.
* Corresponding authors. Tel.: +49 7071 2975247; fax: +49 7071 295137
(M.E.M.), tel.: +44 1865 617585; fax: +44 1865 617575 (U.O.).
E-mail addresses: udo.oppermann@sgc.ox.ac.uk (U. Oppermann), martin.
e.maier@uni-tuebingen.de (M.E. Maier).
One obvious effect of methyl substituents on a macrolactone
core is the resulting conformational control.12 In addition, they
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.11.076

T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
531
11'
Me
OH
O
OH
O
O
OH
1
O
O
O
O
O
3
O
10'
4'
HO
HO
HO
HO
1'
6
zearalenone (1)
HO
6'
5
C
D
O
O
O
O
O
O
O
OH
O
OH
OH
Me
HO
OH
OH
Me
OH
CO2H
O
O
A
B
O
O
E
F
intramolecular aldol
condensation
Figure 3. Reasonable mono-propionate analogues of zearalenone.
Figure 1. Structure and intermediates in the biosynthesis of of zearalenone (1). The
acetate building blocks in A are indicated by red bonds.
reported to comprise high toxicity because of anabolic and oestro-
genic properties. It is also an antibacterial acting chemical.16 It
mimics 17-estradiol, the principal hormone produced by the hu-
man ovary, in binding to oestrogen receptors in mammalian target
cells. Zearalenone is found worldwide as contaminant in cereal
crops and is recognized as endocrine disruptor by having chronic
oestrogenic effects on mammals. A dietary concentration of zearal-
enone as low as 1.0 ppm may lead to hyperestrogenic syndromes
in pigs; higher concentrations can have adverse effects on repro-
duction such as disrupted conception and abortion.17
O
24
OH
15
21
HO
OH
OH
O
O
11
O
OH
OH
7
2
O
OH OH
dictyostatin (2)
erythronolide B (3)
2. Chemistry
O
S
N
We used Mitsunobu macrolactonization to form the macrolac-
tone ring of 5 (Scheme 1). The corresponding seco-acid 6 was cut
further at the vinylic position, leading to iodovinyl benzoate 7
and alkene 8. The alkene may be prepared by addition of a crotyl-
metal species or an equivalent thereof to an aldehyde. In order to
establish the vicinal anti OH/Me pattern at position C60/C50 (zearal-
enone numbering) we used the Marshall–Tamaru reaction,18–20
which relates to the reaction of an aldehyde with a chiral allenyl
zinc reagent.
OH
O
S
O
OH
N
O
OH
O
O
O
epothilone B (4b)
OH
O
epothilone A (4a)
Figure 2. Structures of important polyketides; red = acetate, blue = propionate.
Aldehyde 9 which can be easily prepared from (R)-propylene
oxide,21 was made to react with propargylic mesylate22 10 in pres-
ence of palladium(II) acetate, diethylzinc, and triphenylphosphine
(Scheme 2). The intermediate allenylzinc species then underwent
a SE20 addition to the aldehyde, yielding mainly the anti homoprop-
argylic alcohol 11. Stirring of the reaction mixture for 72 h at
ꢁ20 °C gave the best yield and diastereomeric ratio (9:1, as shown
by NMR, see Supplementary data). The alkyne 11 was hydroge-
nated to the alkene 12 in the presence of Lindlar catalyst (5% Pd
on CaCO3, poisoned with lead). To prevent over-reduction to the
single bond, the Lindlar reduction of alkyne 11 was performed in
may allow the molecule to take advantage of hydrophobic pockets
in the binding site. The compounds epothilone A (4a) and epothi-
lone B (4b) that show anti-cancer activity represent an example
for a natural acetate/propionate pair. Due to the additional methyl
group epothilone B is significantly (approx. 10 times) more active
than epothilone A.13 The significance of the methyl group is also
apparent from the fact that all epothilone derivatives which are
in clinical studies or use are analogues of epothilone B.
We devised the concept of propionate scanning based on the bio-
synthesis of polyketides.14 In general, it comprises the systematic
replacement of all acetates in reasonable positions in polyace-
tate-based natural products with propionate such that one acetate
is replaced in each of the products. A more extended version could
be called methyl-scanning where a methyl group is attached to the
available methylene functions in a macrolactone. However, we feel
the non-propionate analogues would be less validated. We chose
RAL zearalenone as the first compound to apply propionate scan-
ning to. Zearalenone has a relatively simple structure,15 but be-
longs to a very interesting family of natural products of which
many are inhibitors of kinases, ATPases, and chaperones.10 Despite
the lack of obvious similarity with ATP, these compounds have
been shown to bind to the ATP-binding pocket of ATPases includ-
ing kinases. Applied to zearalenone, eight possible analogues can
be conceived from propionate scanning (Fig. 3).
Mitsunobu
OMe
OH
O
10'
CO2H
O
Suzuki
1'
MeO
HO
6'
5'
5'
6'
O
OH
5
6
OMOM
OMe
CO2Me
I
OTBDPS OMOM
+
MeO
7
8
Marshall-
Tamaru
The classical benzolactone, the fungal metabolite zearalenone
(1) was first isolated in 1962 from the fungus Gibberella zeae and
Scheme 1. Retrosynthetic analysis for zearalenone analogue 5.

532
T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
OMs
OTBDPS
O
OMOM
Pd(OAc)2, Ph3P
+
H
B
H
PdCl2(dppf), Ph3As
Et2Zn, THF, —20 °C
(87%)
9-BBN, THF
0 to 23 °C, 12 h
9
10
8
Cs2CO3, H2O/DMF
vinyl iodide 7
G
OTBDPS OH
H
OTBDPS OR
OTBDPS
H2, Lindlar catalyst
EtOAc, 1-octene
OMe
OMe
CO2Me
CO2H
12 R = H
11
MOMCl
i Pr2NEt
(73%)
KOH, H2O/EtOH
reflux (81%)
8 R = MOM
MeO
MeO
OR
OH
Scheme 2. Synthesis of protected homoallyl alcohol
reaction.
8 via Marshall–Tamaru
OMOM
OMOM
16 R = TBDPS
6
HF·pyridine
THF, —30 °C
a solvent mixture containing 1-octene. The hydroxyl function of
the product 12 was protected with MOMCl23 under basic condi-
tions, thus, providing building block 8. The overall yield for the
three step sequence was 65%.
(65%, 2 steps)
17 R = H
MeO
O
18 R = MOM
19 R = H
DEAD,Ph3P
HClconc.
O
The synthesis of the vinyl iodide 7, required for the Suzuki
cross-coupling reaction, started with the known N,N-diethyl-2-for-
myl-benzamide24 13 (Scheme 3). The amide 13 was treated with
HCl (3 N) at 90 °C, resulting in hydrolysis and hemiacetal formation
to provide 3-hydroxy-5,7-dimethoxyphthalide (14). Refluxing the
mixture gave inferior yields. The in situ prepared phthalide was
subsequently methylated under alkylating conditions. Best condi-
tions for the methylation reaction were 40 °C in DMF, giving
methyl 2-formylbenzoate 15 resulting in a good overall yield.
The vinyl iodide 7 could be obtained by Takai olefination25 of alde-
hyde 15. NMR analysis showed that the desired E compound was
obtained in a 5:1:1 ratio over the unwanted Z compound and the
corresponding vinylbenzoate (See Supplementary data). Both by-
products did not interfere in the following Suzuki coupling
reaction.
Building blocks 7 and 8 were combined via a Suzuki cross-cou-
pling reaction (Scheme 4).26,21 Following hydroboration of alkene 8
with 9-BBN (0–23 °C, overnight), the intermediate borane G was
subjected to a palladium-catalyzed coupling reaction with iodosty-
rene 7. Application of a slight excess (1.5 equiv) of 7 in the Suzuki
coupling reaction resulted in a reasonable yield of the alkene 16. To
reach the seco-acid 6, the silyl ether of 16 was cleaved with the
HFꢀpyridine complex in THF, leading to hydroxy ester 17. Saponifi-
cation of the methyl ester with KOH in EtOH/H2O (95:05) under re-
flux resulted in seco-acid 6. For the macrolactone formation
standard Mitsunobu conditions (DEAD, Ph3P, toluene, 0 °C) were
used.27 Thus, a diethyl azodicarboxylate (DEAD) solution was
added slowly to a mixture of seco-acid 6 and triphenylphosphine
in toluene at 0 °C. In order to favour cyclization to 18, the reaction
was performed at relatively high dilution. Cleavage of the MOM
protecting group was accomplished under acidic conditions, thus
providing hydroxy lactone 19.
MeOH
(72%)
toluene, 0 °C
(49%)
MeO
OR
Scheme 4. Synthesis of the lactone 19 via Suzuki cross-coupling and Mitsunobu
macrolactonization as key steps.
Lactone 19 served as the precursor for a few additional zearal-
enone derivatives. Thus, oxidation of 19 with Dess–Martin period-
inane provided macrocyclic ketone 20 (Scheme 5). Treatment of 20
with BBr3 was expected to cleave both methyl ethers.28 However,
under these conditions only the methyl ether ortho to the carboxyl
function was cleaved, thus leading to analogue 21. In order to
cleave both methyl ether functions we found I3Al, prepared in situ,
to be suitable.29 Under these conditions, the zearalenone analogue
5 was obtained in 83% yield.
3. Biology
A family of natural products with high attractivity for the crea-
tion of analogues using propionate scanning are the benzolactones.
Members of this family have a vast spectrum of biological activities
including, for example, protein kinase inhibitors, inhibitors of heat
MeO
O
BBr3, CH2Cl2
–40 °C (74%)
Dess-Martin
O
19
CH2Cl2 (87%)
MeO
O
20
OH
O
O
MeO
O
O
MeO
MeO
O
NEt2
H
MeI, K2CO3
HCl, AcOH
90 °C, 24 h
O
O
21
DMF, 40 °C
(64%)
MeO
MeO
MeO
OH
13
14
OH
O
10'
3
I3Al, TBAI
O
1
MeO
MeO
19
CO2Me
CO2Me
bv
(83%)
0 °C
1'
7'
O
I3CH, CrCl2
HO
5
5'
H
O
THF, 23 °C
(80%)
MeO
3'
5
15
7
I
Scheme 5. Synthesis of various zearalenone analogues from key intermediate 19
Scheme 3. Synthesis of vinyl iodide 7.
via cleavage of the methyl ether functions.

T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
533
shock protein 90 (Hsp90), oestrogen receptor agonists, and antibi-
otics.10,30 As described above, the prototypic benzolactone, zearal-
enone, mimics oestrogen in binding to its receptor, and uses modes
of binding to ATPases similar to that of ATP. Intensive research pro-
grams are ongoing to identify specific kinase inhibitors despite a
high structural similarity within the ATP binding site of these pro-
teins. With an estimated 30% of all cellular proteins becoming tran-
siently phosphorylated the essential role of the protein kinases,
comprising approx. 3% of all expressed genes,31 in cellular signal-
ling pathways is obvious. Their connection to numerous disease
mechanisms including cancer, diabetes and inflammation was
shown in many studies, thus the development of small-molecu-
lar-weight compounds to target inhibition of protein kinases is rec-
ognized as a promising strategy to address various forms of
cancer.32 Hsp90 is an ATP-dependent chaperone that is one of
the most abundant proteins in eukaryotic cells (1–2%). In assisting
non-covalent folding/unfolding of a large number of proteins it
plays a central role for a multitude of functions such as cell cycle
regulation and signal transduction. Pharmacological inhibition of
Hsp90 has been shown to be beneficial in the treatment of multi-
step carcinogenesis.33
It is important to note that the extent of stabilization depends
on the individual properties of the protein; it is dependent on
the entropic and enthalpic components, respectively of the bind-
ing.41 For protein kinases a linear correlation between
DTm and
IC50 is apparent, most likely because of the two-lobe architecture
where the nucleotide (or an inhibitor in its place) exerts most of
its stabilizing effect through ‘locking’ of the two domains to-
gether, which in effect is dominant over the mode of binding
of the inhibitor. For almost all other protein families, however,
the effect of a ligand on the stability of the protein it binds to
is not predictable. On the other hand, however, there is evidence
that for a certain protein the comparison of Tm shifts caused by
different ligands correlate with their affinities. We observed that
the known nanomolar inhibitors of Hsp90, radicicol and geldana-
mycin, caused only a small stabilization of the protein, of 5.1
and 3.2 °C, respectively. In comparison, zearalenone seems to
bind with less affinity to Hsp90. No stabilization was observed
with the analogue 5.
In thermal shift assays CBR1 and other human SDRs such as
CBR3 were screened both with the oxidized and reduced cofactor
(NADP+, NADPH) as well as in the absence of cofactor (Table 2).
In general, consistent with previous finding on a large set of human
oxidoreductases, thermal shifts for SDR proteins are much lower
than for kinases. To the collection of zearalenone analogues, a cou-
ple of structurally related compounds were added to the screening
experiments. Most hits were obtained for CBR1 in the presence of
the cofactor NADP+. Very promising was the obvious selectivity
shown by zearalenone analogue 5 that gave only a distinct Tm shift
with CBR1/NADP+. Taking into account the impaired kinase activity
in comparison to the natural product, this hit seemed auspicious
for further investigations.
Interestingly, binding of inhibitors targeted at protein kinases
to non-ATP binding proteins was observed in
cases. For example, human carbonyl reductase
a
1
number of
(CBR1) is
inhibited by a hydroxyl-substituted analogue of the well-known
Src-family kinase inhibitor.34 CBR1 is a member of a large class
of oxidoreductases, namely short-chain dehydrogenases/reduc-
tases (SDR). CBR1 is one of the major xenobiotic carbonyl
reducing enzymes in humans.35 The outcome of these reactions
is compound and tissue-specific, that is, could lead to more po-
tent or less active compounds. For example, CBR1 catalyzes the
reduction of doxorubicin, a widely used antineoplastic drug, to
the highly cardiotoxic metabolite doxorubicinol. Therefore, inhi-
bition of CBR1 by small-molecular-weight compounds may
proof beneficial in combination therapy during doxorubicin-
based chemotherapy.35
3.1. Substrate Test for CBR1
Stabilization by all of the tested compounds was dependent on
the presence of NADP+. Neither in the presence of the reduced
cofactor, NADPH, nor without cofactor we observed any significant
Tm shift. A likely reason for this effect is that binding of the ligand
requires polar interactions with the positively charged nicotin-
amide ring. However, it could not be ruled out that in presence
of NADPH the compounds would be reduced to a product with
lower affinity, that is, producing a lower stabilizing effect. To test
for the latter possibility, all compounds that caused a Tm shift
(exception: triclosan, which does not comprise any reducible
group) were tested in an activity assay at concentrations of
Within our integrated structural and functional genomics plat-
form directed against human medicinal target classes36 we evalu-
ated the biological potential of the collection of zearalenone
analogues. Binding of compounds to proteins is commonly de-
tected using differential scanning fluorimetry (DSF), a generic ther-
mal-shift (D
Tm) detection assay, as described.37–39 Firstly, we
tested the zearalenone collection (1, 5, 19, 20, 21, including also
geldanamycin and radicicol) against a panel of catalytic domains
of human protein kinases from diverse subfamilies (for the distri-
bution of the selected proteins over the phylogenetic tree of the
human protein kinase family, see Supplementary data, Fig. S1).
As the interaction chart (Table 1) shows, only relatively small ef-
fects were observed. Within this group of compounds, the natural
product zearalenone (1) stabilized lymphocyte-originating kinase
(LOK) and STE20-like kinase (SLK) by little more than 4 °C, com-
monly corresponding to low micromolar inhibition.37 These two
kinases are closely related both structurally and with respect to
promiscuity for kinase inhibitors that are currently either pursued
in clinical trials or approved as drugs.40 In addition to LOK and SLK,
zearalenone also showed stabilization above the threshold (2 °C)
for PIM1, PIM2, PAK5 and GAK. In contrast to these effects, the
additional methyl group, present on the analogue 5 results in a
200 lM, respectively.
However, no CBR1 activity was observed towards the zearale-
none analogue 5. Furthermore, the activity towards the natural
product 1 was below 10% of that towards isatin (Table 3).
3.2. Inhibition of the CBR1-Isatin reaction
We then used the activity of CBR1 towards isatin as an
orthogonal assay to the DSF results, which allows analysis of po-
tential inhibitory properties. In order to study CBR1 inhibition,
we have chosen the CBR1-catalyzed reduction of isatin as refer-
ence reaction (Fig. 4). Isatin has been shown to exhibit various
pharmacological actions and is reduced by CBR1 to 3-hydroxy-
2-oxoindole.
Reductase activity of purified CBR1 was assayed by monitoring
the oxidation of NADPH to NADP+. The fluorescence intensity for
NADPH at 460 nm (after excitation at 355 nm) decreases as the
cofactor is oxidized (Table 4). The reaction rate of the isatin reduc-
tion was determined from linear regression to the data after nor-
malization to the fluorescence increase upon injection (correction
for compound-dependent quenching).
strong decrease in binding affinity: For all but SLK (DDTm
1 °C), Pim1 ( Tm near threshold) and GAK ( Tm similar to that
with zearalenone) the stabilization was lost entirely, that is, the
Tm fell below the threshold of 2 °C. In addition to the effects ob-
ꢂ
D
D
D
served upon addition of the methyl group, the screens also showed
the importance of the aromatic hydroxyl groups for binding to ki-
nases: both dimethoxy analogues (19 and 20) did not bind to any
of the tested proteins.

534
T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
Table 1
Inhibition array of the screened kinases and Hsp90
Yellow fields indicate a Tm shift >2 °C and orange fields of >4 °C, respectively. Blank fields indicate a Tm shift below 2 °C.
Table 3
Table 2
Activity test for compounds with Tm shift >2 °C and additional zearalenone analogues
Stabilization array for human carbonyl reductase 1 (CBR1) and 3 (CBR3)
Compound
Activity compared to isatin (%)
Tm shift (°C)
Wedelolactone
Radicicol
Chrysin
ZEA Analogue 5
Zearalenone (ZEA) (1)
Biochanin A
5.8
—
16.3
—
5.7
9.3
4.7
31.3
—
5.1
4.6
4.6
3.9
3.6
2.7
2.3
<1
Compound
Naringenin
Quercetin
Genistein
ZEA analogue 20
ZOL analogue 19
ZEA analogue 21
<1
<1
1.5
Wedelolactone
2.0
5.1
Geldanamycin,
Radicicol
4.6
3.4
4.6
2.7
2.3
NADP
NADPH
Triclosan
H
N
H
N
Chrysin
O
O
CBR1
Biochanin A
O
OH
3-hydroxy-2-oxoindole
Naringenin
isatin
Dimethoxy ZEA Analogue 19
Dimethoxy ZOL Analogue 20
Mono OMe ZEA Analogue 21
Zearalenone (ZEA) (1)
ZEA Analogue 5
Figure 4. Reduction of isatin by CBR1.
3.6
3.9
Table 4
IC50 values for tested compounds, as well as their Ki values calculated including the
assay concentration of isatin (50
lM)
The threshold for weak binding was set to 2 °C, stabilization effects above the
threshold are colour-coded (Tm shift above, respectively: 2 °C, yellow; 3 °C, ochre;
4 °C, orange; 5 °C, red).
Compound
IC50
(
l
M)
Ki (lM)
Wedelolactone
Radicicol
Triclosan
Chrysin
Biochanin A
Naringenin
3.78 0.9
3.27 2.5
0.40 0.1
0.67 0.3
12.71 8.8
21.48 7.3
0.21 0.03
—
0.60 0.14
0.52 0.40
0.06 0.02
0.11 0.05
2.03 1.40
3.43 1.17
0.04 0.01
—
During the present study, following up on the results from the
DSF screen led to the identification of low micromolar and, in the
case of ZEA analogue 5, even nanomolar CBR1 inhibitors. Interest-
ingly however, not all compounds that stabilized did inhibit the
enzyme strongly. The natural product zearalenone (1) did not lead
to inhibition higher than 50% (see Supplementary data, Fig. S2).
Therefore, no IC50 value could be determined. The binding affinity
of zearalenone to the substrate binding site of CBR1 seems to be
much lower than that of the analogue 5.
ZEA analogue 5
Zearalenone (ZEA) (1)
4. Conclusion
Zearalenone (1) and the analogues 5, 19, 20, and 21 were
screened versus a diverse set of therapeutically relevant human

T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
535
kinases and the chaperone Hsp90. The screening results show that
Acknowledgements
the zearalenone scaffold is likely to serve as a platform for kinase,
Hsp90 or SDR enzyme inhibitor development. Furthermore, the
modification of the underlying framework of the natural product
zearalenone by the synthetic concept called propionate scanning
changed the binding affinity of the compound dramatically for dif-
ferent targets. The incorporation of a propionate building block in
the fourth position with a (S)-methyl group on C50 turned the
resulting zearalenone analogue 5 into a potent (IC50 = 210 nM)
and very specific inhibitor for the carbonyl reductase CBR1. In com-
parison to zearalenone the side effect on inhibition of Hsp90 or ki-
nases is greatly reduced.
Financial support by the Deutsche Forschungsgemeinschaft
(Grant Ma 1012/23-1) and the Fonds der Chemischen Industrie is
gratefully acknowledged. We also thank Graeme Nicholson (Insti-
tute of Organic Chemistry) for measuring the HRMS spectra. T. J.
Zimmermann acknowledges support by the Bayer Science and
Education Foundation and the German National Academic Founda-
tion (Studienstiftung des deutschen Volkes). The Structural
Genomics Consortium is a registered charity (Number 1097737)
and receives funds from the Canadian Institutes for Health Re-
search, the Canadian Foundation for Innovation, Genome Canada
through the Ontario Genomics Institute, GlaxoSmithKline, Karo-
linska Institutet, the Knut and Alice Wallenberg Foundation, the
Ontario Innovation Trust, the Ontario Ministry for Research and
Innovation, Merck and Co., Inc., the Novartis Research Foundation,
the Swedish Agency for Innovation Systems, the Swedish Founda-
tion for Strategic Research and the Wellcome Trust. The authors
would like to thank Dr. Alex Bullock for comments and members
of the SGC PDF group for making recombinant proteins available
for screening.
5. Materials and methods
5.1. Protein expression and purification
All proteins were expressed in Escherichia coli. Detailed proto-
cols for expression and purification of each protein can be down-
loaded from the Structural Genomics Consortium web site
(www.sgc.ox.ac.uk). Either a His6 tag (N- or C-terminal) or a GST
fusion system was used to aid purification. Recombinant proteins
were >95% pure as judged by SDS/PAGE, and protein identity was
confirmed by mass spectrometry.
Supplementary data
Procedures for the syntheses described in Schemes 2–5. Supple-
mentary data associated with this article can be found, in the on-
line version, at doi:10.1016/j.bmc.2008.11.076.
5.2. Compounds
Compounds 5 and 19–21 were synthesized as described in the
Supplementary data. Quercetin, Genistein, Wedelolactone, Gel-
danamycin and Radicicol were purchased from Calbiochem (EMD
Bioscience), and all other compounds from Sigma Aldrich.
References and notes
1. Newman, D. J.; Cragg, G. M. J. Nat. Prod. 2007, 70, 461–477.
2. (a) McArdle, B. M.; Campitelli, M. R.; Quinn, R. J. J. Nat. Prod. 2006, 69, 14–17;
(b) McArdle, B. M.; Quinn, R. J. ChemBioChem 2007, 8, 788–798; (c) Quinn, R. J.;
Carroll, A. R.; Pham, N. B.; Baron, P.; Palframan, M. E.; Suraweera, L.; Pierens, G.
K.; Muresan, S. J. Nat. Prod. 2008, 71, 464–468.
5.3. Thermal shift assays
3. (a) Koch, M. A.; Wittenberg, L.-O.; Basu, S.; Jeyaraj, D. A.; Gourzoulidou, E.;
Reinecke, K.; Odermatt, A.; Waldmann, H. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
16721–16726; (b) Arve, L.; Voigt, T.; Waldmann, H. QSAR Comb. Sci. 2006, 25,
449–456; (c) Umarye, J. D.; Lessmann, T.; Garcia, A. B.; Mamane, V.; Sommer,
S.; Waldmann, H. Chem. Eur. J. 2007, 13, 3305–3319; (d) Leßmann, T.;
Leuenberger, M. G.; Menninger, S.; Lopez-Canet, M.; Müller, O.; Hümmer, S.;
Bormann, J.; Korn, K.; Fava, E.; Zerial, M.; Mayer, T. U.; Waldmann, H. Chem.
Biol. 2007, 14, 443–451.
Differential scanning fluorimetry (DSF) was carried out by using
a Real-Time PCR Mx3005p machine (Stratagene) according to the
procedure described by Niesen et al.38 The protein solutions used
for all fluorescence measurements employed a final protein con-
centration of 2 lM in a buffer consisting of 10 mM HEPES, pH 7.5
and 150 mM NaCl.
4. For recent reviews, see: (a) Hübel, K.; Lessmann, T.; Waldmann, H. Chem. Soc.
Rev. 2008, 37, 1361–1374; (b) Kaiser, M.; Wetzel, S.; Kumar, K.; Waldmann, H.
Cell Mol. Life Sci. 2008, 65, 1186–1201.
5.4. Kinetics of NADPH oxidation
5. For some recent examples, see: (a) Mang, C.; Jakupovic, S.; Schunk, S.; Ambrosi,
H.-D.; Schwarz, O.; Jakupovic, J. J. Comb. Chem. 2006, 8, 268–274; (b) Zinzalla, G.;
Milroy, L.-G.; Ley, S. V. Org. Biomol. Chem. 2006, 4, 1977–2002; (c) Keaney, G. F.;
Johannes, C. W. Tetrahedron Lett. 2007, 48, 5411–5413; (d) Yao, N.; Song, A.;
Wang, X.; Dixon, S.; Lam, K. S. J. Comb. Chem. 2007, 9, 668–676; (e) Schwarz, O.;
Jakupovic, S.; Ambrosi, H.-D.; Haustedt, L. O.; Mang, C.; Müller-Kuhrt, L. J. Comb.
Chem. 2007, 9, 1104–1113; (f) Milroy, L.-G.; Zinzalla, G.; Prencipe, G.; Michel, P.;
Ley, S. V.; Gunaratnam, M.; Beltran, M.; Neidle, S. Angew. Chem. 2007, 119, 2545–
2548; Milroy, L.-G.; Zinzalla, G.; Prencipe, G.; Michel, P.; Ley, S. V.; Gunaratnam,
M.; Beltran, M.; Neidle, S. Angew. Chem., Int. Ed. 2007, 46, 2493–2496.
6. For a summary, see: Hruby, V. J. Nat. Rev. Drug Discovery 2002, 1, 847–858.
7. For a recent example, see: Nam, J.; Shin, D.; Rew, Y.; Boger, D. L. J. Am. Chem.
Soc. 2007, 129, 8747–8755.
The assay was performed in wells of a 384-well low-volume
microtiter plate (Corning); the final reaction volume was 30 lL.
Oxidation of NADPH to NADP was monitored from the decrease
in fluorescence at 460 nm (excitation: 355 nm) after addition of
200 nM protein and 200 lM NADPH to the compound(s) dissolved
in 50 mM sodium phosphate, pH 6.8, 1 mM MgCl2 (injector inte-
grated into plate reader: Polarstar Omega, BMG Labtech). The reac-
tion rate was determined from linear regression to the data after
normalization to the fluorescence increase upon injection (correc-
tion for compound-dependent quenching). Ability of compounds to
function as CBR1 substrates was tested in triplicates at concentra-
8. For reviews, see (a) Hof, F.; Diederich, F. Chem. Commun. 2004, 477–480; (b)
Müller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881–1886; (c) Zürcher, M.;
Diederich, F. J. Org. Chem. 2008, 73, 4345–4361.
9. Baumann, S.; Schoof, S.; Harkal, S. D.; Arndt, H.-D. J. Am. Chem. Soc. 2008, 130,
5664–5666 and references cited therein.
tions of 200
l
M, respectively. Inhibition was determined by titra-
M, 200 M).
tion in presence of two concentrations of isatin (50
l
l
10. Winssinger, N.; Barluenga, S. Chem. Commun. 2007, 22–36.
Background activities (in absence of protein) at each substrate con-
centration were subtracted, and data fitted to a sigmoidal dose-re-
sponse model to determine the IC50 (Graphpad Prism 5). Inhibition
constants, Ki, were calculated according to the equation:
11. Gaffoor, I.; Trail, F. Appl. Environ. Microbiol. 2006, 72, 1793–1799.
12. (a) Hoffmann, R. W. Chem. Rev. 1989, 89, 1841–1860; (b) Hoffmann, R. W.;
Stahl, M.; Schopfer, U.; Frenking, G. Chem. Eur. J. 1998, 4, 559–566.
13. (a) Nicolaou, K. C.; Ritzén, A.; Namoto, K. Chem. Commun. 2001, 1523–1535; (b)
Altmann, K.-H.; Cachoux, F.; Caravatti, G.; Isarno, T.; Wartmann, M. ARKIVOC
2006, 131–141.
14. (a) Staunton, J.; Weissman, K. Nat. Prod. Rev. 2001, 18, 380–416; (b) Fischbach,
M. A.; Walsh, C. T. Chem. Rev. 2006, 106, 3468–3496.
15. For total syntheses of zearalenone, see: (a) Takahashi, T.; Ikeda, H.; Tsuji, J.
Tetrahedron Lett. 1981, 22, 1363–1366; (b) Kalivretenos, A.; Stille, J. K.;
Hegedus, L. S. J. Org. Chem. 1991, 56, 2883–2894; (c) Fürstner, A.; Thiel, O. R.;
IC50
Ki ¼
;
s
1 þ
Km
where s, concentration of substrate isatin; Km, Michaelis constant
for isatin (45.9 10.2 M).
l

536
T. J. Zimmermann et al. / Bioorg. Med. Chem. 17 (2009) 530–536
Kindler, N.; Bartkowska, B. J. Org. Chem. 2000, 65, 7990–7995; (d) Nicolaou, K. 29. Liang, Q.; Sun, Y.; Yu, B.; She, X.; Pan, X. J. Org. Chem. 2007, 72, 9846–
C.; Winssinger, N.; Pastor, J.; Murphy, F. Angew. Chem. 1998, 110, 2677–2680;;
Nicolaou, K. C.; Winssinger, N.; Pastor, J.; Murphy, F. Angew. Chem., Int. Ed.
1998, 37, 2534–2537.
9849.
30. Yet, L. Chem. Rev. 2003, 103, 4283–4306.
31. Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.; Sudarsanam, S. Science
2002, 298, 1912–1916. 1933–1934.
16. Stob, M.; Baldwin, R. S.; Tuite, J.; Andrews, F. N.; Gillette, K. G. Nature 1962, 196,
1318.
32. Karaman, M. W.; Herrgard, S.; Treiber, D. K.; Gallant, P.; Atteridge, C. E.;
Campbell, B. T.; Chan, K. W.; Ciceri, P.; Davis, M. I.; Edeen, P. T.; Faraoni, R.;
Floyd, M.; Hunt, J. P.; Lockhart, D. J.; Milanov, Z. V.; Morrison, M. J.; Pallares, G.;
Patel, H. K.; Pritchard, S.; Wodicka, L. M.; Zarrinkar, P. P. Nat. Biotechnol. 2008,
26, 127–132.
33. Proisy, N.; Sharp, S. Y.; Boxall, K.; Connelly, S.; Roe, S. M.; Prodromou, C.;
Slawin, A. M. Z.; Pearl, L. H.; Workman, P.; Moody, C. J. Chem. Biol. 2006, 13,
1203–1215.
34. Tanaka, M.; Bateman, R.; Rauh, D.; Vaisberg, E.; Ramachandani, S.; Zhangg, C.;
Hansen, K. C.; Burlingame, A. L.; Trautman, J. K.; Shokat, K. M.; Adams, C. L. PLoS
Biol. 2005, 3, 764–776.
35. For a review, see: Oppermann, U. Annu. Rev. Pharmacol. Toxicol. 2007, 47, 293–
322.
36. Gileadi, O.; Knapp, S.; Lee, W. H.; Marsden, B. D.; Muller, S.; Niesen, F. H.;
Kavanagh, K. L.; Ball, L. J.; von Delft, F.; Doyle, D. A.; Oppermann, U. C. T.;
Sundstrom, M. J. Struct. Funct. Genomics 2007, 8, 107–119.
37. Vedadi, M.; Niesen, F. H.; Allali-Hassani, A.; Fedorov, O. Y.; Finerty, P. J., Jr.;
Wasney, G. A.; Yeung, R.; Arrowsmith, C.; Ball, L. J.; Berglund, H.; Hui, R.;
Marsden, B. D.; Nordlund, P.; Sundstrom, M.; Weigelt, J.; Edwards, A. M. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 15835–15840.
38. Niesen, F. H.; Berglund, H.; Vedadi, M. Nat. Protoc. 2007, 2, 2212–2221.
39. Bullock, A. N.; Debreczeni, J. E.; Fedorov, O. Y.; Nelson, A.; Marsden, B. D.;
Knapp, S. J. Med. Chem. 2005, 48, 7604–7614.
17. Bennett, J. W.; Klich, M. Clin. Microbiol. Rev. 2003, 16, 497–516.
18. (a) Tamaru, Y.; Goto, S.; Tanaka, A.; Shimizu, M.; Kimura, M. Angew. Chem.
1996, 108, 962–963; Tamaru, Y.; Goto, S.; Tanaka, A.; Shimizu, M.; Kimura, M.
Angew. Chem., Int. Ed. 1996, 35, 878–880; (b) Marshall, J. A.; Adams, N. D. J. Org.
Chem. 1998, 63, 3812–3813; (c) Marshall, J. A.; Adams, N. D. J. Org. Chem. 1999,
64, 5201–5204; (d) Marshall, J. A.; Eidam, P.; Eidam, H. S. J. Org. Chem. 2006, 71,
4840–4844; (e) Marshall, J. A.; Eidam, P.; Schenck, H. Org. Synth. 2007, 84, 120–
128.
19. For a review, see: Marshall, J. A. Chem. Rev. 2000, 100, 3163–3185.
20. (a) Prusov, E.; Röhm, H.; Maier, M. E. Org. Lett. 2006, 8, 1025–1028; (b) Jägel, J.;
Maier, M. E. Synlett 2006, 693–696.
21. Khartulyari, A. S.; Manmohan, K.; Maier, M. Org. Lett. 2006, 8, 5833–5836.
22. Marshall, J. A.; Chobanian, H. Org. Synth. 2005, 82, 43–54.
23. Reggelin, M.; Doerr, S. Synlett 2004, 1117.
24. Brimble, M. A.; Houghton, S. I.; Woodgate, P. D. Tetrahedron 2006, 63, 880–887.
25. Takai, K.; Nitta, K.; Utimoto, K. J. Am. Chem. Soc. 1986, 108, 7408–7410.
26. For some reviews, see: (a) Chemler, S. R.; Trauner, D.; Danishefsky, S. J. Angew.
Chem. 2001, 113, 4676–4701; Chemler, S. R.; Trauner, D.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2001, 40, 4544–4568; (b) Bellina, F.; Carpita, A.; Rossi, R. Synthesis
2004, 2419–2440; (c) Nicolaou, K. C.; Bulger, P. G.; Sarlah, D. Angew. Chem. 2005,
117, 4516–4563; Nicolaou, K. C.; Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed.
2005, 44, 4442–4489; (d) Suzuki, A. Chem. Commun. 2005, 4759–4763.
27. Hughes, D. L. Org. React. 1992, 42, 335–656.
40. Fedorov, O.; Marsden, B.; Pogacic, V.; Rellos, P.; Mueller, S.; Bullock, A. N.;
Schwaller, J.; Sundstroem, M.; Knapp, S. Proc. Natl. Acad. Sci. U.S.A. 2007, 104,
20523–20528.
28. (a) Taub, D.; Girotra, N. N.; Hoffsommer, R. D., Jr.; Kuo, C.-H.; Slates, H. L.;
Weber, S.; Wendler, N. L. Tetrahedron 1968, 24, 2443–2461; (b) Hitchcock, S. A.;
Pattenden, G. J. Chem. Soc., Perkin Trans. 1 1992, 1323–1328.
41. Holdgate, G. A.; Ward, W. H. Drug Discov. Today 2005, 10, 1543–1550.