The role of fluorine substitution in biphenyl methylene imidazole-type CYP17 inhibitors for the treatment of prostate carcinoma

Article abstract of DOI:10.1002/cmdc.201000065

It has been established that the growth of most prostate carcinomas depends on androgen stimulation. The inhibition of cytochrome P450-17 (CYP17) to block androgen biosynthesis is therefore regarded as a promising approach to therapy. Based on our previou

Full text of DOI:10.1002/cmdc.201000065

DOI: 10.1002/cmdc.201000065
The Role of Fluorine Substitution in Biphenyl Methylene
Imidazole-Type CYP17 Inhibitors for the Treatment of
Prostate Carcinoma
Qingzhong Hu,[a] Matthias Negri,[a] Sureyya Olgen,[a, b] and Rolf W. Hartmann*[a]
It has been established that the growth of most prostate carci-
nomas depends on androgen stimulation. The inhibition of cy-
tochrome P450-17 (CYP17) to block androgen biosynthesis is
therefore regarded as a promising approach to therapy. Based
on our previously identified lead compound Ref1, a series of
fluorine-substituted biphenyl methylene imidazoles were de-
signed, synthesized, and evaluated as CYP17 inhibitors to eluci-
date the influence of fluorine on in vitro and in vivo activity. It
was found that meta-fluoro substitution at the C ring improved
activity, whereas ortho substitution decreased potency. Dock-
ing studies performed with our human CYP17 homology
model suggest the presence of multipolar interactions be-
tween fluorine and Arg109, Lys231, His235, and Glu305. As ex-
pected, introduction of fluorine also prolonged the half-life in
plasma. The SARs obtained confirm the reliability of the pro-
tein model; compound 9 (IC50 =131 nm) was identified as a
strong CYP17 inhibitor, showing potent activity in rat, high bio-
availability, and a long plasma half-life: 12.8 h.
Introduction
As the most common malignancy in elder men, prostate carci-
noma is a major cause of death.[1] It has been established that
the growth of up to 80% of prostate carcinomas depends on
androgen stimulation. Therefore, the segregation of tumor
cells from androgen hormones would effectively prevent
cancer cell proliferation. Because >90% of testosterone is pro-
duced in the testes, orchidectomy or treatment with gonado-
tropin-releasing hormone (GnRH) analogues[2] (chemical castra-
tion) are applied in the clinic. As this therapy has no effect on
the minor amounts of androgen produced in the adrenal
glands, androgen receptor antagonists are used in conjunction.
This is the current standard therapy for prostate carcinoma,
the so-called “combined androgen blockade” (CAB).[3] However,
CAB often leads to resistance, which can be associated with
androgen receptor mutations. The mutated androgen receptor
recognizes antagonists and glucocorticoids as agonists, ulti-
mately resulting in the collapse of CAB therapy.[4]
and progesterone to dehydroepiandrosterone (DHEA) and an-
drostenedione, respectively. DHEA can be transformed into an-
drostenedione by 3b-hydroxysteroid dehydrogenase (3b-HSD),
followed by its subsequent conversion into the most potent
androgen, dihydrotestosterone (DHT), in androgen target cells
through two enzymatic steps catalyzed by 17b-HSD1 or 17b-
HSD3 and steroid 5a-reductase (5aR). Thus, inhibition of
CYP17 could block androgen production in testes as well as
adrenal tissue. Furthermore, targeting genetically stable
healthy tissue instead of cancer cells would circumvent resist-
ance caused by mutation.
Ketoconazole (Figure 1), an antimycotic agent that shows
nonselective inhibition of CYP17, is the first medication which
has been used clinically in the treatment of prostate carcino-
ma. Although withdrawn because of side effects, ketoconazole
shows good curative properties,[8] which demonstrates the fea-
sibility of prostate carcinoma treatment via CYP17 inhibition.
Since then, in mimicry of the physiological substrates, many
steroidal CYP17 inhibitors were synthesized by others[9] as well
as our research group;[10] among these, abiraterone (Figure 1)
recently entered phase II clinical trials. However, the affinity of
The shortcomings of CAB warrant a more promising alterna-
tive: total blockage of androgen biosynthesis, which means
the inhibition of cytochrome P450-17 (17a-hydroxylase-17, 20-
lyase, CYP17). CYP17 is one of six CYP enzymes involved in ste-
roid biosynthesis. Like all CYP enzymes, CYP17 consists of a
heme and an apoprotein moiety. Although all potent inhibitors
interfere with the heme (which is common to all CYP enzymes)
by complexing its central iron ion, it is nevertheless possible to
selectively inhibit these enzymes, as has been demonstrated
with CYP19 (aromatase, estrogen synthase)[5] and CYP11B2 (al-
dosterone synthase).[6] Whereas CYP19 inhibitors are already in
clinical use,[5a] the first highly potent and selective CYP11B2 in-
hibitors were just recently identified,[6] some of which are ex-
tremely selective.
[a] Q. Hu,+ M. Negri,+ Dr. S. Olgen, Prof. Dr. R. W. Hartmann
Pharmaceutical and Medicinal Chemistry, Saarland University
P.O. Box 151150 & Helmholtz Institute for Pharmaceutical Research
Saarland (HIPS), 66041 Saarbrꢀcken (Germany)
Fax: (+49)681-302-70308
E-mail: rwh@mx.uni-saarland.de
Homepage: http://www.PharmMedChem.de
[b] Dr. S. Olgen
Present address: University of Ankara, Faculty of Pharmacy
Tandogan, Ankara 06100 (Turkey)
[+] These authors contributed equally to this work.
CYP17, located in both testicular and adrenal tissues,[7] is the
key enzyme that catalyzes the conversion of pregnenolone
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000065.
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
899

R. W. Hartmann et al.
MED
the heme iron center, recognized as the main anchor
point for nonsteroidal CYP inhibitors (first noted for
CYP19 inhibitors[5b–c]), polar interactions between the
Ref1 fluorine atom and the guanidinium side chain
of Arg109 and the amine group of Lys231 were ob-
served and considered important for binding affini-
ty[12g] (Figure 2). The A ring is presumably stabilized
by a strong T-shaped arene quadrupole interaction
Figure 1. Typical CYP17 inhibitors.
steroidal compounds toward steroid receptors, which often re-
sults in side effects regardless of whether the compound in
question acts as an agonist or antagonist, prompted us to de-
velop nonsteroidal CYP17 inhibitors.[11,12]
Our research group has reported a series of biphenyl meth-
ylene imidazoles as potent CYP17 inhibitors.[12] A promising
lead compound 1-[1-(4’-fluorobiphenyl-4-yl)propyl]-1H-imidazo-
le[12g] (Ref1, Figure 1) was identified in the optimization pro-
cess. In the present study this compound was further modified
to increase its potency, selectivity, and to improve its pharma-
cokinetic properties. Fluorine is known to be able to form mul-
tipolar interactions with several amino acids,[13e–l] and it can
also enhance metabolic stability; for these reasons, the biphen-
yl core was substituted with additional fluorine atoms, leading
to compounds 1–22. Exchanging the 1-imidazolyl group with a
5-imidazolyl moiety while maintaining 4-fluorophenyl as the
A ring, compounds 23–26 were subsequently obtained. Fur-
thermore, besides determination of inhibitory activities toward
human CYP17 in vitro, selected compounds were examined for
their ability to decrease plasma testosterone concentrations
and for their pharmacokinetic properties in rats. Moreover,
computational investigations were performed: molecular dock-
ing studies using our homology model of human CYP17[12e] to
elucidate the enzyme–inhibitor interactions and quantum me-
chanical studies to explore the influence of fluorine substitu-
tion on the potency and pharmacokinetic properties of this
type of CYP17 inhibitor.
Figure 2. CYP17 residues surrounding Ref1: Arg109 and Lys231 interact with
the para-fluoro group on the A ring. More amino acids around the A or
C rings might form additional multipolar interactions with further fluoro sub-
stituents.
with Phe114, a conformationally flexible residue responsible for
dividing the CYP17 active site into two lobes. Furthermore, ad-
ditional amino acid residues such as Asn, Arg, and Gln were
observed to be proximal to the A or C rings, and might pro-
vide the potential for additional fluorine substituents to form
multipolar interactions with HꢀX (for X=N, O, S),[13h–l] back-
bone C=O groups (in an orthogonal manner),[13e–f] or even with
HꢀCa.[13c,g] Consequently, the following strategies to increase
the inhibitory potency and metabolic stability of Ref1 were ap-
plied: a) shifting the fluorine to other positions in the A ring,
and b) introduction of additional fluorine atoms on the A or
C rings to identify new interaction areas. Because these modifi-
cations change the molecular electrostatic potentials (MEPs) of
the compounds, the protein pocket surrounding the C ring
was scrutinized to identify potential interaction areas influ-
enced by MEP variations, especially the backbone p systems of
amino acids such as Gly301–Ala302.
Drug Design
Fluorine, as the most electronegative atom, forms strong CꢀF
bonds and has therefore been widely used to prevent unde-
sired metabolism. In addition to increased metabolic stability,
fluorine can also improve other pharmacokinetic properties by
influencing pKa, elevating lipophilicity, and decreasing plasma
protein binding.[13] Recently, multipolar interactions between
fluorine and some amino acid residues responsible for en-
hanced binding potency have also been reported.[13e–l] Based
on these findings, the concepts of fluorophilicity and fluoro- Results and Discussion
phobicity in protein active sites have been discussed,[13c,d] and
Chemistry
systematic fluorine scans have been recommended for drug
discovery and lead optimization efforts.
The syntheses of compounds 1–26 are shown in Schemes 1, 2,
We found that fluorine substitution at the para position of
the A ring could significantly increase the inhibitory potency of
biphenyl methylene imidazole-type CYP17 inhibitors, resulting
in compound Ref1 (IC50 =345 nm).[12g] In addition to complex
formation between a heterocyclic nitrogen atom of Ref1 and
and 3. For 1-imidazole analogues 1–22, a general synthetic
strategy was employed: ketone or aldehyde intermediates
were obtained by Suzuki coupling (Method C) from the corre-
sponding bromides and boronic acids.[14] They were subse-
quently converted into the alcohols by reduction with NaBH4
900
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910

Fluorine Substitution in CYP17 Inhibitors
Scheme 1. Compounds were synthesized from the corresponding a or b intermediates unless otherwise indicated. Reagents and conditions: a) Method C:
Pd(PPh3)4, Na2CO3, toluene, reflux, 6 h; b) Method D: NaBH4, MeOH; c) Method B: EtMgBr, THF; d) Method E: CDI, THF, reflux, 4 days; e) Method A: CDI, NMP,
reflux, 4 h.
(Method D) or Grignard reaction (Method B). The alcohol inter-
mediates reacted with 1,1’-carbonyldiimidazole (CDI) to give
racemic mixtures of the desired products, which were not sep-
arated into their enantiomers. By varying the reaction condi-
tions such as solvent and temperature, different products were
obtained in this SN2 reaction. After holding at reflux in N-
methyl-2-pyrrolidone (NMP) for 4 h (Method A),[15] biphenyl
methylene imidazoles were obtained, whereas boiling in THF
for four days (Method E) gave imidazole-1-carboxylic acid bi-
phenyl esters as major products. Distinguishing between these
Scheme 3. Reagents and conditions: a) 1. Method C: 4-fluoroboronic acid,
two products is easy, as the chemical shift values of imidazole
Pd(PPh3)4, Na2CO3, toluene, reflux, 6 h; 2. pyridinium·HCl, MeOH, 608C, 4 h;
b) HCl/iPrOH, 808C, 2 h; c) Pd(OH)2, EtOH, THF, H2, room temperature, 3 h.
2-H in biphenyl methylene imidazoles are ~7.6 ppm, whereas
they are >8.2 ppm in imidazole-1-carboxylic acid biphenyl
esters.[15] Moreover, the proton chemical shift of CꢀH is
5.1 ppm in the biphenyl methylene imidazoles, whereas in the
imidazole-1-carboxylic acid biphenyl esters it is 5.9 ppm. Addi-
tionally, the strong carbonyl signal in the IR spectra contributes
to the identification of the latter class of compounds. For the
synthesis of the 5-imidazole analogues 23–26 (Schemes 2 and
3), triphenylmethyl (Trt) was employed as the protecting group
for imidazole. The alcohol intermediates, commercially avail-
able or obtained by reaction of imidazolyl lithium with alde-
hyde, underwent elimination of the H2O group under acidic
conditions to give the corresponding isopropylidene product,
which was subsequently saturated by hydrogenation of the
double bond.
In vitro activity
CYP17 inhibition of all compounds was evaluated by using the
50000 g sediment after homogenization of E. coli expressing
human CYP17 as well as cytochrome P450 reductase.[12d] The
assay was run with progesterone as substrate and NADPH as
cofactor. Separation of substrate and product was carried out
by HPLC with UV detection.[16] IC50 values are presented in rela-
tive to ketoconazole and abiraterone in Tables 1 and 2.
It is striking that in the series of 1-imidazole compounds (1–
22) a sharp structure–activity relationship is observed. The ana-
logues can be divided into three classes based on fluorine sub-
stitution on the C ring: without fluoro, meta-fluoro, and ortho-
fluoro (positions relative to the A ring, Table 1). It is apparent
that in the class of compounds lacking fluorine at the C ring,
Scheme 2. Reagents and conditions: a) Method C: 4-fluoroboronic acid,
Pd(PPh3)4, Na2CO3, toluene, reflux, 6 h; b) 1. imidazole, nBuLi, TrtCl, THF, 08C,
2 h; 2. nBuLi, TBDMSCl, THF, 08C, 2 h; 3. nBuli, 23b, THF, room temperature,
8 h; c) pyridinium·HCl, MeOH, 608C, 4 h.
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
901

R. W. Hartmann et al.
MED
Table 1. Inhibition of CYP17 by compounds 1–22.
Compd
R1
R2
H
R3
Et
IC50 [nm][a]
Compd
R1
R2
R3
IC50 [nm][a]
Compd
19
R1
R2
R3
Et
IC50 [nm][a]
Ref1
4’-F
345
9
4’-F
4’-F
4’-F
4’-F
m-F
m-F
m-F
m-F
Et
Et
H
131
4643
2110
4’-F
o-F
657
10
11
12
20
21
22
4’-F
o-F
o-F
o-F
H
Et
H
2800
825
H
>5000
1
2
3
4
5
6
7
4’-F, 2’-Me
3’,4’-diF
3’,4’-diF
2’,4’-diF
2’,4’-diF
H
H
H
H
H
H
H
H
Et
Et
Et
Et
Et
Me
Et
Et
951
803
>5000
985
>5000
>10000
956
>5000
2780
13
14
15
16
3’,4’-diF
3’,4’-diF
2’,4’-diF
2’,4’-diF
m-F
m-F
m-F
m-F
Et
H
Et
H
305
>5000
381
3’,4’-diF
2’,4’-diF
>5000
>5000
2’,4’-diF, 3’-MeO
2’,5’-diF
17
18
2’,5’-diF
2’,5’-diF
m-F
m-F
Et
H
364
1640
8
2’,5’-diF
KTZ[b]
ABT[b]
72
[a] Compound concentration required to give 50% inhibition; values represent the mean ꢁ10% deviation of at least three experiments. [b] KTZ: ketocona-
zole, ABT: abiraterone.
additional substitution at the A ring by fluoro, methyl, or me-
thoxy groups (1, 2, 4, 6, and 7), does not enhance the inhibito-
ry activity relative to Ref1 (IC50 =345 nm). For example, extra
fluorine substitution at the meta or ortho positions of the
A ring, resulting in 3’,4’-difluoro analogue 2 and 2’,4’-difluoro
analogue 4, decreased activity somewhat (IC50: 803 and
985 nm, respectively). Interestingly, 2’,5’-difluoro substitution
(compound 7) resulted in similar inhibitory potency (IC50 =
956 nm) as observed with the 2’,4’-difluoro analogue.
Moreover, all imidazole-1-carboxylic acid biphenyl esters (3,
5, 8, 10, and 12), which have been shown by HPLC to be
stable in incubation buffer, were found to be inactive. This is
probably due to the fact that these compounds are too large
to fit into the binding pocket. Notably, the ethyl substituent at
the methylene bridge is important, as previously reported;[12g]
that is, ethyl substitution results in compounds that are more
potent than the corresponding unsubstituted or methyl-substi-
tuted analogues.
Importantly, the introduction of fluorine at the meta position
of the C ring significantly increases inhibitory potency. Com-
pound 9 has an IC50 value of 131 nm, being threefold more
potent than Ref1. Similar improvements can be found for the
other compounds in this class (13, 15, and 17; IC50 values
~350 nm), which are threefold more potent than the corre-
sponding compounds without fluoro group substitution at the
C ring (2, 4, and 7, respectively; IC50 values ~900 nm). More-
over, the analogue with only one para-fluoro substituent at
the A ring (compound 9) is more potent than other analogues
with multiple fluorine atoms on the A ring, as mentioned
above. A similar potency ranking is observed for polyfluorinat-
ed analogues in this compound class as well: the 3’,4’-difluoro
compound 13 (IC50 =305 nm) is more potent than the others:
for example, the 2’,5’-difluoro compound 15 (IC50 =381 nm)
and the 2’,4’-difluoro compound 17 (IC50 =364 nm).
In contrast, ortho-fluoro substitution at the C ring decreases
activity. Compound 19 (IC50 =657 nm) is less potent than the
corresponding analogues with meta-fluoro groups (compound
9, IC50 =131 nm) or those lacking fluorine (Ref1, IC50 =345 nm).
However, compound 19 exhibited stronger inhibition than
compound 21, which bears two fluorine substituents at the 3’
and 4’ positions, as expected.
Furthermore, 5-imidazole was employed instead of 1-imida-
zole, leading to compounds 23–26 as shown in Table 2. It is
apparent that compounds with a hydroxy group substituted at
the methylene bridge are inactive. Nonetheless, isopropyl sub-
stitution resulted in an active compound (26, IC50 =502 nm)
which was, however, less potent than Ref1. Surprisingly, com-
Table 2. Inhibition of CYP17 by compounds 23–26.
Compd
R5
R6
IC50 [nm][a]
23
24
25
OH
OH
H
iPr
@10000
>5000
159
isopropylidene
26
H
iPr
502
2780
72
KTZ[b]
ABT[b]
[a] Compound concentration required to give 50% inhibition; values rep-
resent the mean ꢁ10% deviation of at least three experiments. [b] KTZ:
ketoconazole, ABT: abiraterone.
902
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910

Fluorine Substitution in CYP17 Inhibitors
pound 25, with an isopropylidene substitution, turned out to
be very potent (IC50 =159 nm).
and the R enantiomers into BM-ABT. However, as the intersec-
tion of both areas, BM1 is a better area for both enantiomers
to bind.
Finally, inhibition values of the most active compounds 9,
13, 15, 17, and 25 toward the hepatic enzyme CYP3A4 were
also determined because of its important role in drug metabo-
lism and drug–drug interactions. It turned out that the com-
pounds tested showed only marginal inhibition (~50% at
10 mm), clearly lower than that shown by ketoconazole (98% at
10 mm).
It was observed for BM1 that a conjugated scaffold oriented
almost parallel to the I-helix is one of the key factors for high
activity, as previously described.[12g] According to our docking
studies this extended p system clearly interacts not only with
the p system of the amino acid backbone in the I-helix (i.e.,
Gly301, Ala302, Gly303, and Val304), but also with Phe114,
which is oriented perpendicular to the A ring (Figure 3, BM1)
to form a quadrupole–quadrupole interaction. More important-
ly, fluorine in the molecule showed profound influence on the
affinity for the enzyme. This can be explained by the clear in-
teractions of the para-fluoro substituent with Arg109, Lys231,
and His235 (Figure 3). When the fluorine atoms were shifted to
other positions in the A ring resulting in 2’,5’-difluoro ana-
logues, a decrease in activity can be observed. Clearly, the in-
teractions mentioned above can no longer be maintained.
These observations nicely validate the reliability of our protein
model. Furthermore, the elevated activity of C ring meta-
fluoro-substituted compounds can be explained by a multipo-
lar interaction between F and the NꢀH group of Glu305
(Figure 3) stabilizing the p–p interaction between the C ring
and the backbone p system of Gly301–Ala302.
Computational studies
Docking
Compounds Ref1, 9, 11, 13, 15, 17, and 19, both enantiomers
if present, were docked into the homology model of human
CYP17[12e] by means of two commercial docking software pack-
ages, GOLD_v. 4.0[17] and FlexX 3.1.3,[18] in order to elucidate
their binding into the active site of the enzyme. Docking with
GOLD was performed with both the scoring functions Gold-
Score and ChemScore, while FlexX was used with the FlexX-
Pharm module, with the iron center of the heme chosen as
pharmacophoric constraint. A clustering with ACIAP[19] of all
the docking poses of the three procedures resulted in one
main statistically predominant binding mode (BM1)[12e] and
two minor representative clusters. Interestingly, each of the
latter two clusters consisted solely of either the S (BM2)[12e] or
the R (BM-ABT)[12h] enantiomers (Figure 3). This finding indi-
cates that, together with the necessary perpendicular interac-
tion angle between imidazole N and heme Fe to ensure suffi-
cient coordination, the orientation of the hydrophobic pocket,
which is occupied by the substituent on the methylene bridge,
limits the pose distribution of different enantiomers. This geo-
metrical restriction forces the S enantiomers into the BM2 area
Additionally, a further flexible docking run was performed
with GOLD v. 4.0–GoldScore, with the side chains of Phe114,
Arg109, Lys231, Asn202, Glu305, and Ile371 being set as freely
rotatable. Very similar results were observed as those for the
docking runs with the rigid side chains (data not shown).
MEP maps
To gain insight into which physicochemical parameters might
influence biological activity, the charge density distribution
was considered, and the molecular electrostatic potentials of
selected compounds were determined. The geometry of these
compounds (Ref1, 2, 9, 13, 15, 17, and 25) was optimized in
the gas phase at the B3LYP/6-311++G** (d,p)[20] level of densi-
ty functional theory (DFT) by means of Gaussian 03.[21] MEPs of
electron density were plotted for every compound with Gauss-
View 3.09.[22]
The introduction of a second fluorine or the shift of fluorine
into another position on the A ring always leads to a decrease
in A ring electron density, which can be clearly observed in the
various MEP maps of compounds 9, 13, 15, 17, and 25 (Fig-
ure 4A). This observation correlates with the different electron-
withdrawing effects of the fluorine atom in the meta or para
position (s values of +0.337 and +0.067, respectively). This
decrease in electron density weakens the T-shaped interaction
of the A ring with Phe114, and consequently decreases the in-
hibitory potency of the compound.
The lower inhibitory potency of compounds with ortho-
fluoro substitution on the C ring (compounds 19 and 21)
might be due to the adverse effect that the fluorine exerts on
the overall charge density of the biphenyl system by deform-
ing the conjugated p–p system and by concentrating the elec-
trons on the fluorine. However, for meta-fluoro compounds 9,
Figure 3. Presentation of the three main binding modes exemplified by
compounds 9 (magenta, BM1), 15 (red, BM2) and 13 (blue, BM-ABT). Here,
interacting residues and the tertiary structure (ribbon) of the active site are
shown. Polar interactions are marked with red solid lines, whereas p–p
stacking and metal complexation are indicated with cyan dotted lines. This
image was generated with MOE (http://www.chemcomp.com).
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
903

R. W. Hartmann et al.
MED
Figure 4. A) Structure, CYP17 inhibitory activity, and MEP maps (bottom and top view) of compounds 9 and 13. For compounds 15 and 17, nearly the same
MEP maps (not shown) as for compound 13 were observed. B) MEP maps of compound 25, highlighting the side views of the conjugated p–p system en-
hanced by the isopropylidene substituent on the methylene bridge. The electrostatic potential surfaces shown in A) and B) were plotted with GaussView 3.0
in a range of ꢀ6 to +12 kcalmolꢀ1. C) Transparent color-coded front and back view of the electrostatic potential maps of the imidazole groups of both Ref1
and compound 9, in a range of ꢀ15.7 to +0 kcalmolꢀ1, illustrating the influence exerted on the electron distribution of the imidazole by fluorine substituents
on the C ring.
Table 3. Pharmacokinetic properties of selected compounds.[a]
13, 15, and 17, it seems that fluorine induces an increase in
electron distribution on the imidazole ring, as visualized in the
[f]
Compd
t
= z [h][b] tmax [h][c] Cmax [nm][d] C24h [nm][e] AUC0–1 [nmhꢀ1
]
1
2
MEP maps of compounds Ref1 and 9 (Figure 4C), which is as-
sociated with an augmented inhibitory potency of the com-
pounds. An analogous phenomenon was observed for com-
pound 25, in which the isopropylidene group strengthens the
p–p system and extends the conjugation over the whole mole-
cule (Figure 4B).
9
13
15
17
Ref1
ABT
12.8
6.1
4.2
2.2
10.0
1.6
2.0
8.0
6.0
1.0
6.0
2.0
3473
3496
2539
458
11729
1694
1528
1173
556
187
4732
253
80448
56114
35091
2055
252297
11488
[a] Compounds Ref1, 9, 13, 15, and 17 were applied at a dose of
50 mgkgꢀ1; abiraterone (ABT) was administrated as abiraterone acetate
(56 mgkgꢀ1, equivalent to ABT 50 mgkgꢀ1). Five to six intact adult male
Wistar rats were employed for each treatment group; each sample was
tested three times. [b] Terminal half-life. [c] Time of maximal concentra-
tion. [d] Maximal concentration. [e] Concentration at 24 h. [f] Area under
the curve.
In vivo activity
The in vivo evaluation of the most potent compounds 9, 13,
15, and 17, including the ability to decrease plasma testoster-
one concentrations (Figure 5) and the determination of phar-
macokinetic properties (Table 3), was performed in male Wistar
rats after oral application. Abiraterone, which was administered
in its acetate form to improve oral absorption, and Ref1 were
used as reference compounds. The plasma concentrations of
testosterone were determined using an ELISA assay, and
plasma drug concentrations were measured by LC–MS. In the
case of abiraterone acetate, only the signals of free abiraterone
were monitored, as the acetate is inactive as a CYP17 inhibitor.
As can be seen in Figure 5, all compounds significantly de-
creased the plasma testosterone concentration. It is striking at
each time point investigated that all nonsteroidal compounds,
which were less active in vitro, exhibited higher activities
in vivo than abiraterone. After 24 h, compound 9 and Ref1 still
showed strong inhibitory activity, compounds 13 and 17
showed almost no inhibition, and compound 15 and abirater-
one exhibited an increase in testosterone levels above control
at this time point, probably caused by feedback stimulation.
As expected, this activity profile correlates to the pharmacoki-
netic properties of the compounds. Abiraterone exhibited a
plasma half-life of only 1.6 h, while the half-lives of other com-
pounds are much longer (10 h for Ref1 and 12.8 h for 9). Ac-
cordingly, the AUC values of the test compounds, except for
compound 17, are higher than that of abiraterone, indicating
better bioavailability. Interestingly, the introduction of an addi-
904
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910

Fluorine Substitution in CYP17 Inhibitors
interacting amino acid residues significantly increased the
binding affinities relative to the parent compound. The charge
distribution difference on both the A and C rings indicates p–p
stacking (i.e., with both Phe114 and Gly301–Ala302), hydro-
phobic, and van der Waals interactions as determinants for ac-
tivity. Furthermore, it was demonstrated once again that fluo-
rine substituted in an appropriate position, like in the C ring,
prolongs the plasma half-life; in an unsuitable position, howev-
1
er, it decreases the t = z value of the parent compound. This
2
phenomenon might be due to a decrease in metabolic stability
and is an interesting subject for future investigations.
Finally, after the modification, compound 9 was identified as
a strong CYP17 inhibitor showing potent activity in vivo, high
bioavailability, and a long plasma half-life. Thus, compound 9
appears to be an optimal candidate, which, after further struc-
tural optimization, could be the first nonsteroidal CYP17 inhibi-
tor to be applied clinically. CYP17 inhibitors are expected to be
superior to the GnRH analogues in current use as mentioned
above, because they decrease not only testicular but also adre-
nal androgen formation. Nevertheless, PC treatment could be
further optimized by combining CYP17 inhibitors with inhibi-
tors of androgen activation to DHT, catalyzed by 17b-HSD1[23]
and/or 17b-HSD3, as well as 5aR.[24]
Figure 5. Reduction of plasma testosterone concentrations in rats by select-
ed compounds. Compounds Ref1, 9, 13, 15, and 17 were applied at a dose
of 50 mgkgꢀ1; abiraterone (ABT) was administrated as abiraterone acetate
(56 mgkgꢀ1, equivalent to abiraterone at 50 mgkgꢀ1). Five to six intact adult
male Wistar rats were employed for each treatment group; each sample was
tested three times. The average plasma testosterone concentrations
(1.97 ngmLꢀ1) at pre-treatment time points (ꢀ1, ꢀ0.5 and 0 h) were set at
100%. The values shown are relative to the pre-treatment value.
tional fluorine into the C ring of Ref1 prolongs plasma half-life
(compound 9), whereas introduction of further fluorine atoms
into the A ring significantly decreases half-life values (13, 15,
Experimental Section
1
and 17; t = : 2–6 h). This is probably due to the electron-with-
CYP17 preparation and assays
2
drawing effects of multiple fluoro group substitution. Fluorine
Human CYP17 was expressed in E. coli (co-expressing human
CYP17 and cytochrome P450 reductase), and the assay was per-
formed as previously described.[12d,16]
atoms significantly weaken the adjacent aromatic CꢀH bonds,
making them more vulnerable to nucleophilic attack.[13m]
Conclusions
Inhibition of hepatic CYP enzyme
Herein we report the design, synthesis, and bioactivity evalua-
tion of a series of fluorine-substituted biphenyl methylene imi-
dazoles as CYP17 inhibitors. Fluorine substitution showed a
profound influence on the in vitro and in vivo activities, as well
as the pharmacokinetic profiles.
The recombinantly expressed enzyme from baculovirus-infected
insect microsome (Supersome) was used, and the manufacturer’s
instructions (http://www.gentest.com) were followed.
In vivo studies
We observed that fluorine at the meta position of the C ring
increases activity relative to unsubstituted analogues, whereas
ortho substitution decreases potency. Compounds bearing flu-
orine at the A ring consistently follow the same activity rank-
ing: 4’-F > 3’,4’-diF > 2’,5’-diF ꢂ 2’,4’-diF. Although the com-
pounds are racemates, it is not likely that one enantiomer
would be much more potent than the other, because both S
and R enantiomers predominately adopt BM1 according to our
docking studies. Close inspection of the docking poses reveals
that the biphenyl moieties of both enantiomers are placed in
the same area, interacting with Arg109 and Lys231, and the
positioning of their imidazolyl methylene groups exhibits only
a slight difference. The conformational flexibility of the imida-
zolyl methylene group allows the ethyl substituents of both
enantiomers to orient toward the hydrophobic pocket that is
opposite to the I-helix and delimited by Ile371 and Ala367.
Moreover, the biological data were well deciphered by docking
studies and MEP mapping of selected compounds. The multi-
polar interactions between the C ring fluorine substituents and
The in vivo tests were performed with intact adult male Wistar rats
(Harlan Winkelmann, Germany), five to six for each treatment
group. These rats were cannulated with silicone tubing via the
right jugular vein. Compounds 9, 13, 15, and 17 were applied p.o.
at doses of 50 mgkgꢀ1, while abiraterone was administrated as the
acetate at 56 mgkgꢀ1 (equivalent to abiraterone at 50 mgkgꢀ1).
The concentrations of testosterone in the rat plasma were deter-
mined by ELISA (EIA-1559) from DRG Instruments according to the
manufacturer’s instructions. The plasma drug levels were measured
by LC–MS. Non-compartmental pharmacokinetic analysis of con-
centration versus time was performed for each compound on the
mean plasma level using a validated computer program (PK solu-
tion 2 software, Summit Research Services, Montrose, CO, USA).
Plasma concentrations below the limit of detection were assigned
a value of zero.
Chemistry
General: Melting points were determined on a Mettler FP1 melting
point apparatus and are uncorrected. IR spectra were recorded
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
905

R. W. Hartmann et al.
MED
neat on a Bruker Vector 33FTIR spectrometer. 1H and 13C NMR
spectra were measured on a Bruker DRX-500 (500 MHz). Chemical
shifts (d) are given in parts per million (ppm), and (CH3)4Si was
used as an internal standard for spectra obtained. All coupling con-
stants (J) are given in Hz. ESI (electrospray ionization) mass spectra
were determined on a TSQ Quantum (Thermo Electron Corpora-
tion) instrument. High-resolution mass spectra were measured
using an LTQ Orbitrap (Thermo Electron Corporation) with positive
ESI. The purities of the final compounds were determined by a Sur-
veyor LC system and were >98%. Column chromatography was
performed using silica gel 60 (50–200 mm), and reaction progress
was monitored by TLC on Alugram SIL G/UV254 (Macherey–Nagel).
Boronic acids and bromoaryl compounds used as starting materials
were commercially obtained (CombiBlocks, Chempur, Aldrich,
Acros, etc).
d=1.75 (d, J=7.5 Hz, 3H, CH3), 3.91 (s, 3H, OCH3), 5.18–5.31 (q, J=
7.5 Hz, 1H, CH), 6.82–6.83 (m, 1H), 6.84–6.86 (m, 1H), 6.89–6.93 (m,
1H), 6.97–6.99 (m, 1H), 7.09–7.10 (m, 2H), 7.34–7.36 (m, 2H), 7.50–
7.51 ppm (m, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=21.1 (CH3),
56.7 (CH), 62.5 (OCH3), 112.3, 117.7, 123.9, 125.6, 127.8, 128.4, 134.1,
136.1, 141.6 ppm; HRMS (ESI): calcd for C18H17F2N2O [M+H]+:
315.1309, found: 315.1297; MS (ESI): m/z=315 [M+H]+.
1-[1-(2’,5’-Difluorobiphenyl-4-yl)propyl]-1H-imidazole (7). Synthe-
sized according to Method A using 7a (0.27 g, 1.09 mmol) and CDI
(0.34 g, 2.06 mmol); yield: 0.13 g (41%); colorless oil; Rf =0.38
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.98 (t, J=
7.3 Hz, 3H, CH3), 2.26 (q, J=7.3, 7.6 Hz, 2H, CH2), 5.05 (t, J=7.6 Hz,
1H, CH), 6.97–7.01 (m, 2H), 7.08–7.12 (m, 3H), 7.25–7.27 (m, 2H),
7.49–7.51 (m, 2H), 7.62–7.65 ppm (m, 1H, Im-2H); 13C NMR (CDCl3,
125 MHz): d=11.3 (CH3), 29.6 (CH2), 63.1 (CH), 116.8, 118.4, 126.5,
130.5, 135.4, 136.9, 140.0, 158.8, 160.7 ppm; HRMS (ESI): calcd for
C18H17F2N2 [M+H]+: 299.1360, found: 299.1351; MS (ESI): m/z=299
[M+H]+.
Method A: CDI reaction in NMP. CDI (5 equiv) was added to a so-
lution of the corresponding alcohol (1 equiv) in NMP or CH3CN
(10 mLmmolꢀ1). The solution was then heated at reflux for 4–18 h.
After cooling to room temperature, it was diluted with H2O (30 mL)
and extracted with EtOAc (3ꢁ10 mL). The combined organic
phases were washed with brine, dried over MgSO4 and evaporated
under reduced pressure. The desired product was then purified by
chromatography on silica gel.
1-[1-(3,4’-Difluorobiphenyl-4-yl)propyl]-1H-imidazole (9). Synthe-
sized according to Method A using 9a (0.38 g, 1.53 mmol) and CDI
(0.47 g, 2.88 mmol); yield: 0.08 g (17%); yellow oil; Rf =0.38
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.99 (t, J=
7.3 Hz, 3H, CH3), 2.21–2.31 (m, 2H, CH2), 5.38 (t, J=7.5 Hz, 1H, CH),
7.01 (s, 1H), 7.08 (s, 1H), 7.10–7.14 (m, 2H), 7.19–7.25 (m, 2H),
7.27–7.29 (m, 1H), 7.40–7.51 (m, 2H), 7.64 ppm (s, 1H, Im-2H);
13C NMR (CDCl3, 125 MHz): d=11.3 (CH3), 28.6 (CH2), 56.4 (CH),
114.5, 116.3, 118.0, 123.7, 126.6, 127.3, 128.9, 129.5, 130.9, 135.0,
136.5, 161.7, 163.4 ppm; 19F NMR (CDCl3, 400 MHz): d=ꢀ114.19 (s,
1F), ꢀ117.98 ppm (s, 1F); HRMS (ESI): calcd for C18H17F2N2 [M+H]+:
299.1360, found: 299.1352; MS (ESI): m/z=299 [M+H]+.
1-[1-(4’-Fluoro-2’-methylbiphenyl-4-yl)propyl]-1H-imidazole (1).
Synthesized according to Method A using 1a (0.15 g, 0.61 mmol)
and CDI (0.20 g, 1.23 mmol); yield: 0.06 g (32%); colorless oil; Rf =
0.27 (CH2Cl2/MeOH, 95:5); 1H NMR (DMSO, 500 MHz): d=0.98 (t,
J=7.3 Hz, 3H, CH3), 2.23 (s, 3H, CH3), 2.24–2.30 (m, 2H, CH2), 5.07
(t, J=7.6 Hz, 1H, CH), 6.90–6.93 (m, 1H), 6.95–6.97 (m, 1H), 7.01 (s,
1H), 7.11 (s, 1H), 7.14 (dd, J=6.0, 8.4 Hz, 1H), 7.22–7.26 (m, 4H),
7.67 ppm (s, 1H); 13C NMR ([D6]DMSO, 125 MHz): d=11.1 (CH3),
20.5 (CH3), 28.6 (CH2), 63.2 (CH), 112.5, 112.6, 116.7, 116.9, 126.3,
129.7, 131.1, 131.1, 137.0, 137.6, 137.7, 138.9, 140.9, 161.1,
163.0 ppm; HRMS (ESI): calcd for C19H20FN2 [M+H]+: 295.1611,
found: 295.1607; MS (ESI): m/z=295 [M+H]+.
1-[(3,4’-Difluorobiphenyl-4-yl)methyl]-1H-imidazole (11). Synthe-
sized according to Method A using 11 a (0.28 g, 1.27 mmol) and
CDI (0.41 g, 2.54 mmol); yield: 0.11 g (31%); colorless oil; Rf =0.53
(EtOAc/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=5.20 (s, 2H,
CH2), 6.98 (brs, 1H), 7.10 (brs, 1H), 7.11–7.16 (m, 3H), 7.26–7.31 (m,
2H), 7.50 (dd, J=5.4, 8.8 Hz, 2H), 7.64 ppm (brs, 1H, Im-2H);
13C NMR (CDCl3, 125 MHz): d=44.4 (CH2), 114.2, 115.9, 119.2, 122.2,
123.1, 128.6, 129.6, 129.9, 135.3, 137.3, 142.9, 160.6, 162.9 ppm;
HRMS (ESI): calcd for C16H13F2N2 [M+H]+: 271.1047, found:
271.1039; MS (ESI): m/z=271 [M+H]+.
1-[1-(3’,4’-Difluorobiphenyl-4-yl)propyl]-1H-imidazole (2). Synthe-
sized according to Method A using 2a (0.27 g, 1.09 mmol) and CDI
(0.34 g, 2.06 mmol); yield: 0.09 g (27%); colorless oil; Rf =0.29
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.98 (t, J=
7.3 Hz, 3H, CH3), 2.23–2.31 (q, J=7.3, 7.6 Hz, 2H, CH2), 5.06 (t, J=
7.6 Hz, 1H, CH), 6.97–7.02 (m, 1H), 7.10–7.14 (m, 1H), 7.18–7.28 (m,
4H), 7.23–7.37 (m, 1H), 7.49 (m, 2H), 7.62–7.66 ppm (m, 1H, Im-
2H); 13C NMR (CDCl3, 125 MHz): d=11.3 (CH3), 29.6 (CH2), 63.0 (CH),
116.3, 117.5, 118.7, 123.2, 127.3, 130.5, 136.5, 137.4, 140.5,
151.3 ppm; HRMS (ESI): calcd for C18H17F2N2 [M+H]+: 299.1360,
found: 299.1363; MS (ESI): m/z=299 [M+H]+.
1-[1-(3,3’,4’-Trifluorobiphenyl-4-yl)propyl]-1H-imidazole (13). Syn-
thesized according to Method A using 13a (0.40 g, 1.70 mmol) and
CDI (0.53 g, 3.20 mmol); yield: 0.11 g (20%); colorless oil; Rf =0.35
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.99 (t, J=
7.3 Hz, 3H, CH3), 2.22–2.33 (m, 2H, CH2), 5.39 (t, J=7.6 Hz, 1H, CH),
7.01 (s, 1H), 7.09 (s, 1H), 7.20–7.29 ( m, 5H), 7.31–7.35 (m, 1H),
7.64 ppm (s, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=11.3 (CH3),
28.6 (CH2), 56.3 (CH), 114.5, 116.2, 118.4, 123.9, 127.5, 128.3, 130.5,
136.7, 149.4, 161.7 ppm; 19F NMR (CDCl3, 400 MHz): d=ꢀ117.56 (s,
1-[1-(2’,4’-Difluorobiphenyl-4-yl)propyl]-1H-imidazole (4). Synthe-
sized according to Method A using 4a (0.27 g, 1.09 mmol) and CDI
(0.34 g, 2.06 mmol); yield: 0.12 g (36%); colorless oil; Rf =0.34
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.85 (t, J=
7.3 Hz, 3H, CH3), 2.12–2.18 (m, 2H, CH2), 4.96 (t, J=7.6 Hz, 1H, CH),
6.77–6.85 (m, 2H), 6.87–6.90 (m, 1H), 6.99–7.02 (m, 1H), 7.13–7.15
(m, 2H), 7.23–7.28 (m, 1H), 7.35–7.37 (m, 2H), 7.62–7.68 ppm (m,
1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=11.1 (CH3), 29.3 (CH2),
63.3 (CH), 104.5, 112.7, 117.8, 118.6, 124.6, 126.3, 130.2, 131.4,
135.7, 136.6, 140.3, 163.5 ppm; HRMS (ESI): calcd for C18H17F2N2
[M+H]+: 299.1360, found: 299.1352; MS (ESI): m/z=299 [M+H]+.
3
3
1F), ꢀ136.80 (d, JFF =ꢀ20.7, 1F), ꢀ138.64 ppm (d, JFF =ꢀ20.7,
1F); HRMS (ESI): calcd for C18H16F3N2 [M+H]+: 317.1266, found:
317.1257; MS (ESI): m/z=317 [M+H]+.
1-[(3,3’,4’-Trifluorobiphenyl-4-yl)methyl]-1H-imidazole (14). Syn-
thesized according to Method A using 14a (0.35 g, 1.67 mmol) and
CDI (0.53 g, 3.20 mmol); yield: 0.10 g (23%); white powder, mp:
1
60–618C; Rf =0.26 (CH2Cl2/MeOH, 95:5); H NMR (CDCl3, 500 MHz):
1-[1-(2’,4’-Difluoro-3’-methoxybiphenyl-4-yl)ethyl]-1H-imidazole
(6). Synthesized according to Method A using 6a (0.30 g,
1.09 mmol) and CDI (0.34 g, 2.06 mmol); yield: 0.34 g (71%);
d=5.19 (s, 2H, CH2), 6.97 (s, 1H), 7.10 (s, 1H), 7.11–7.14 (m, 2H),
7.20–7.28 (m, 3H), 7.31–7.35 (m, 1H), 7.59 ppm (s, 1H, Im-2H);
13C NMR (CDCl3, 125 MHz) 44.3 (CH2), 114.9, 116.4, 118.6, 119.3,
123.9, 129.5, 136.8, 137.3, 141.4, 149.3, 151.7, 161.4 ppm; HRMS
1
yellow oil; Rf =0.30 (CH2Cl2/MeOH, 95:5); H NMR (CDCl3, 500 MHz):
906
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910

Fluorine Substitution in CYP17 Inhibitors
(ESI): calcd for C16H12F3N2 [M+H]+: 289.0953, found: 289.0944; MS
(ESI): m/z=289 [M+H]+.
1-(2,4’-Difluorobiphenyl-4-ylmethyl)-1H-imidazole (20). Synthe-
sized according to Method A using 20a (0.10 g, 0.44 mmol) and
CDI (0.40 g, 2.46 mmol); yield: 0.05 g (42%); white solid, mp: 93–
1
1-[1-(2’,3,4’-Trifluorobiphenyl-4-yl)propyl]-1H-imidazole (15). Syn-
thesized according to Method A using 15a (0.27 g, 1.14 mmol) and
CDI (0.35 g, 2.10 mmol); yield: 0.09 g (24%); colorless oil; Rf =0.37
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.99 (t, J=
7.3 Hz, 3H, CH3), 2.23–2.32 (m, 2H, CH2), 5.40 (t, J=7.6 Hz, 1H, CH),
6.89–6.97 (m, 2H), 7.02 (s, 1H), 7.08 (s, 1H), 7.19–7.27 (m, 3H),
7.34–7.38 (m, 1H), 7.64 ppm (s, 1H, Im-2H); 13C NMR (CDCl3,
125 MHz): d=11.3 (CH3), 27.7 (CH2), 56.2 (CH), 104.5, 112.7, 116.9,
117.7, 125.4, 127.0, 130.8, 136.5, 137.9, 160.3, 161.2 ppm; 19F NMR
948C; Rf =0.52 (EtOAc/MeOH, 95:5); H NMR (CDCl3, 500 MHz): d=
5.15 (s, 2H, CH2), 6.90–6.98 (m, 3H), 7.11–7.13 (m, 2H), 7.25–7.27
(m, 3H), 7.36–7.39 (m, 1H), 7.60 ppm (s, 1H, Im-2H); 13C NMR
(CDCl3, 125 MHz) 50.0 (CH2), 114.9, 115.1, 115.5, 119.3, 123.0, 129.9,
130.6, 130.7, 130.9, 131.2, 137.6, 158.8, 160.8, 161.6, 163.6 ppm;
HRMS (ESI): calcd for C16H13F2N2 [M+H]+: 271.1047, found:
271.1045; MS (ESI): m/z=271 [M+H]+.
1-[1-(2,3’,4’-Trifluorobiphenyl-4-yl)propyl]-1H-imidazole (21). Syn-
thesized according to Method A using 21a (0.28 g, 1.03 mmol) and
CDI (0.34 g, 2.06 mmol); yield: 0.09 g (29%); light-brown oil; Rf =
4
4
(CDCl3, 400 MHz): d=ꢀ109.86 (d, JFF =8.0, 1F), ꢀ113.15 (d, JFF =
8.0, 1F), ꢀ118.03 ppm (s, 1F); HRMS (ESI): calcd for C18H16F3N2
[M+H]+: 317.1266, found: 317.1257; MS (ESI): m/z=317 [M+H]+.
1
0.53 (EtOAc/MeOH, 95:5); H NMR (CDCl3, 500 MHz): d=0.97 (t, J=
7.4 Hz, 3H), 2.20–2.28 (m, 2H), 5.04 (t, J=7.6 Hz, 1H), 6.96–6.99 (m,
2H), 7.03 (dd, J=1.9, 7.9 Hz, 1H), 7.11 (brs, 1H), 7.18–7.23 (m, 2H),
7.31–7.34 (m, 1H), 7.35 (t, J=8.0 Hz, 1H), 7.64 ppm (brs, 1H, Im-
2H); 13C NMR (CDCl3, 125 MHz): d=10.0 (CH3), 27.5 (CH2), 61.5 (CH),
113.5, 116.4, 116.5, 117.0, 121.6, 124.0, 125.8, 128.7, 129.9, 130.8,
135.3, 141.5, 149.0, 149.2, 158.6 ppm; HRMS (ESI): calcd for
C18H16F3N2 [M+H]+: 317.1266, found: 317.1257; MS (ESI): m/z=317
[M+H]+.
1-[(2’,3,4’-Trifluorobiphenyl-4-yl)methyl]-1H-imidazole (16). Syn-
thesized according to Method A using 16a (0.18 g, 0.76 mmol) and
CDI (0.24 g, 1.51 mmol); yield: 0.09 g (42%); orange oil; Rf =0.52
(EtOAc/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=5.20 (s, 2H,
CH2), 6.89–6.97 (m, 2H), 6.98 (brs, 1H), 7.10 (brs, 1H), 7.12 (dd, J=
7.9, 8.2 Hz, 1H), 7.24–7.27 (m, 2H), 7.36 (ddd, J=6.4, 8.5, 8.8 Hz,
1H), 7.61 ppm (brs, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=44.3
(CH2), 104.6, 111.8, 116.2, 119.2, 122.8, 123.3, 125.1, 129.4, 129.8,
131.1, 137.4, 159.2, 159.6, 161.2, 162.7 ppm; HRMS (ESI): calcd for
C16H12F3N2 [M+H]+: 289.0953, found: 289.0952; MS (ESI): m/z=289
[M+H]+.
1-(2,2’,4’-Trifluorobiphenyl-4-ylmethyl)-1H-imidazole (22). Syn-
thesized according to Method A using 22a (0.10 g, 0.39 mmol) and
CDI (0.40 g, 2.46 mmol); yield: 0.06 g (54%); white solid, mp: 112–
1
113 8C; Rf =0.47 (EtOAc/MeOH, 95:5); H NMR (CDCl3, 500 MHz): d=
5.21 (s, 2H, CH2), 6.90–6.98 (m, 4H), 7.02 (d, J=7.8 Hz, 1H), 7.16 (s,
1H), 7.33–7.35 (m, 2H), 7.85 ppm (s, 1H, Im-2H); 13C NMR (CDCl3,
125 MHz): d=52.2 (CH2), 104.3, 111.5, 114.8, 119.4, 122.8, 129.1,
132.2, 132.2, 137.3, 138.3, 158.9, 161.0, 161.9, 162.0, 164.1 ppm;
HRMS (ESI): calcd for C16H12F3N2 [M+H]+: 289.0953, found:
289.0957; MS (ESI): m/z=289 [M+H]+.
1-[1-(2’,3,5’-Trifluorobiphenyl-4-yl)propyl]-1H-imidazole (17). Syn-
thesized according to Method A using 17a (0.40 g, 1.70 mmol) and
CDI (0.53 g, 3.20 mmol); yield: 0.14 g (26%); colorless oil; Rf =0.38
(CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.99 (t, J=
7.3 Hz, 3H, CH3), 2.23–2.32 (m, 2H, CH2), 5.40 (t, J=7.6 Hz, 1H, CH),
7.00–7.04 (m, 2H), 7.08–7.14 (m, 2H), 7.20–7.24 (m, 1H), 7.26–7.30
(m, 2H), 7.64 ppm (s, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=
11.3 (CH3), 27.6 (CH2), 56.3 (CH), 116.7, 117.9, 125.5, 127.0, 130.3,
136.7, 158.5, 159.7, 161.4 ppm; 19F NMR (CDCl3, 400 MHz): d=
Method B: Grignard reaction. Under exclusion of air and moisture,
a solution of EtMgBr (1.0m, 1.2 equiv) in THF was added dropwise
to
a solution of the aldehyde or ketone (1 equiv) in THF
(12 mLmmolꢀ1). The mixture was stirred overnight at room tem-
perature. Then EtOAc (10 mL) and H2O (10 mL) were added, and
the organic phase was separated. The organic phase was extracted
with H2O and brine, dried over Na2SO4, and evaporated under re-
duced pressure. The crude products were purified by flash chroma-
tography on silica gel.
5
5
ꢀ117.81 (s, 1F), ꢀ118.39 (d, JFF =17.7, 1F), ꢀ123.67 ppm (d, JFF =
17.7, 1F); HRMS (ESI): calcd for C18H16F3N2 [M+H]+: 317.1266,
found: 317.1257; MS (ESI): m/z=317 [M+H]+.
1-[(2’,3,5’-Trifluorobiphenyl-4-yl)methyl]-1H-imidazole (18). Syn-
thesized according to Method A using 18a (0.17 g, 0.71 mmol) and
CDI (0.23 g, 1.43 mmol); yield: 0.09 g (45%); light-yellow solid, mp:
125–1268C; Rf =0.48 (EtOAc/MeOH, 95:5); 1H NMR (CDCl3,
500 MHz): d=5.21 (s, 2H, CH2), 6.98 (brs, 1H), 7.00–7.05 (m, 1H),
7.08–7.16 (m, 4H), 7.28–7.31 (m, 2H), 7.64 ppm (brs, 1H, Im-2H);
13C NMR (CDCl3, 125 MHz): d=44.4 (CH2), 116.0, 116.4, 116.6, 117.4,
119.3, 125.1, 123.4, 129.5, 129.8, 137.1, 137.4, 156.5, 154.6, 158.7,
160.2 ppm; HRMS (ESI): calcd for C16H12F3N2 [M+H]+: 289.0953,
found: 289.0955; MS (ESI): m/z=289 [M+H]+.
Method C: Suzuki coupling. The corresponding brominated aro-
matic compound (1 equiv) was dissolved in toluene (7 mLmmolꢀ1),
and an aqueous solution of 2.0m Na2CO3 (3.2 mLmmolꢀ1) and an
ethanolic solution (3.2 mLmmolꢀ1) of the corresponding boronic
acid (1.5–2.0 equiv) were added. The mixture was deoxygenated
under reduced pressure and flushed with N2. After repeating this
cycle several times Pd(PPh3)4 (4 mol%) was added, and the result-
ing suspension was heated at reflux for 8 h. After cooling, EtOAc
(10 mL) and H2O (10 mL) were added, and the organic phase was
separated. The aqueous phase was extracted with EtOAc (2ꢁ
10 mL). The combined organic phases were washed with brine,
dried over Na2SO4, filtered over a short plug of Celite, and evapo-
rated under reduced pressure. The compounds were purified by
flash chromatography on silica gel.
1-[1-(2,4’-Difluorobiphenyl-4-yl)propyl]-1H-imidazole (19). Syn-
thesized according to Method A using 19a (0.31 g, 1.22 mmol) and
CDI (0.40 g, 2.46 mmol); yield: 0.11 g (31%); colorless oil; Rf =0.56
(EtOAc/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.97 (t, J=
7.3 Hz, 3H), 2.21–2.27 (m, 2H), 5.04 (t, J=7.7 Hz, 1H), 6.96–6.99 (m,
2H), 7.02 (dd, J=1.9, 7.9 Hz, 1H), 7.11 (brs, 1H), 7.12 (t, J=8.6 Hz,
2H), 7.37 (t, J=7.9 Hz, 1H), 7.47 (ddd, J=1.3, 5.4, 8.8 Hz, 1H),
7.64 ppm (brs, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=11.0
(CH3), 28.4 (CH2), 62.5 (CH), 114.3, 115.4, 117.5, 122.4, 127.8, 129.6,
130.5, 131.0, 136.3, 141.8, 141.9, 159.6, 162.5 ppm; HRMS (ESI):
calcd for C18H17F2N2 [M+H]+: 299.1360, found: 299.1351; MS (ESI):
m/z=299 [M+H]+.
1-(4’-Fluorobiphenyl-4-yl)-1-(1H-imidazol-5-yl)-2-methylpropan-
1-ol (24). Synthesized according to Method C using 1-(4-bromo-
phenyl)-2-methyl-1-(1-triphenylmethyl-1H-imidazol-5-yl)propan-1-ol
(0.50 g, 0.93 mmol) and 4-fluorophenylboronic acid (0.23 mg,
1.63 mmol). After workup, the crude was stirred with pyridi-
nium·HCl (0.17 g, 1.5 mmol) in MeOH (40 mL) at 608C for 4 h. The
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
907

R. W. Hartmann et al.
MED
reaction was then quenched by adding saturated NaHCO3(aq)
(10 mL). The phases separated after the addition of EtOAc (20 mL).
The aqueous phase was extracted with EtOAc (2ꢁ20 mL), the com-
bined organic extracts were dried over Na2SO4 and concentrated
under reduced pressure. The crude product was purified by flash
chromatography; yield: 0.21 g (73%); colorless oil; Rf =0.17
(EtOAc/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz): d=0.82 (d, J=
6.8 Hz, 3H), 0.98 (d, J=6.8 Hz, 3H), 2.63–2.65 (m, 1H), 7.01 (s, 1H),
7.10 (dd, J=8.7, 8.8 Hz, 2H), 7.49 (d, J=8.5 Hz, 2H), 7.53 (dd, J=
5.4, 8.7 Hz, 2H), 7.58 (s, 1H), 7.62 ppm (d, J=8.3 Hz, 2H); 13C NMR
(CDCl3, 125 MHz): d=17.3, 60.4, 77.7, 115.5, 115.6, 126.2, 126.4,
128.4, 128.5, 133.9, 136.9, 136.9, 138.3, 144.8, 161.4, 163.3 ppm;
HRMS (ESI): calcd for C19H20FN2O [M+H]+: 311.1560, found:
311.1559; MS (ESI): m/z=311 [M+H]+.
7.09–7.15 (m, 3H), 7.46–7.48 (m, 3H), 7.52–7.55 (m, 2H), 8.23 ppm
(s, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=9.9 (CH3), 28.9 (CH2),
82.2 (CH), 115.5, 116.7, 117.6, 126.9, 127.4, 129.4, 130.2, 135.3,
136.9, 138.4, 148.0, 154.7, 156.7, 157.8, 159.4 ppm; HRMS (ESI):
calcd for C19H17F2N2O2 [M+H]+: 343.1258, found: 343.1250; MS
(ESI): m/z=343 [M+H]+.
Imidazole-1-carboxylic acid 1-(3,4’-difluorobiphenyl-4-yl)propyl
ester (10). Synthesized according to Method E using 7a (0.19 g,
0.75 mmol) and CDI (0.24 g, 1.50 mmol); yield: 0.10 g (69%); color-
less oil; Rf =0.29 (CH2Cl2/MeOH, 95:5); 1H NMR (CDCl3, 500 MHz):
d=1.01 (t, J=14.8 Hz, 3H, CH3), 2.02–2.20 (q, J=14.8, 15.5 Hz, 2H,
CH2), 5.88 (t, J=15.5 Hz, 1H, CH), 6.14 (s, 1H), 7.26–7.28 (m, 2H),
7.19–7.25 (m, 1H), 7.34–7.35 (m, 1H), 7.42–7.48 (m, 2H), 7.50–7.53
(m, 2H), 8.22 ppm (s, 1H, Im-2H); 13C NMR (CDCl3, 125 MHz): d=
9.7 (CH3), 28.1 (CH2), 82.2 (CH), 114.3, 114.4, 115.8, 116.0, 117.2,
122.9, 128.1, 128.7, 135.3, 137.0, 159.3, 161.3, 161.9, 163.9,
171.1 ppm; HRMS (ESI): calcd for C19H17F2N2O2 [M+H]+: 343.1258,
found: 343.1250; MS (ESI): m/z=343 [M+H]+.
Method D: Reduction with NaBH4. NaBH4 (2 equiv) was added to
an ice-cooled solution of the corresponding aldehyde or ketone
(1 equiv) in MeOH (5 mLmmolꢀ1). The resulting mixture was then
heated at reflux for 30 min. After cooling to room temperature, the
solvent was distilled off under reduced pressure. H2O (10 mL) was
then added, and the resulting mixture was extracted with EtOAc
(3ꢁ10 mL). The combined organic phases were washed with brine,
dried over MgSO4 and evaporated under reduced pressure. Then
the desired product was purified by chromatography on silica gel.
Imidazole-1-carboxylic acid 3,4’-difluorobiphenyl-4-ylmethyl
ester (12). Synthesized according to Method E using 11 a (0.26 g,
1.20 mmol) and CDI (0.39 g, 2.40 mmol); yield: 0.09 g (28%); white
solid, mp: 220–2218C; Rf =0.29 (CH2Cl2/MeOH, 95:5); 1H NMR
(CDCl3, 500 MHz): d=5.33 (s, 2H, CH2), 7.07 (s, 1H), 7.12–7.15 (t, J=
17.4 Hz, 2H), 7.20 (s, 1H), 7.29 (d, J=11.0 Hz, 1H), 7.33–7.35 (m,
2H), 7.49–7.52 (m, 2H), 8.32 ppm(s, 1H, Im-2H); 13C NMR (CDCl3,
125 MHz): d=45.3 (CH2), 114.3, 115.9, 116.0, 121.1, 123.3, 126.7,
128.6, 128.7, 130.6, 130.7, 143.5, 143.6, 161.8, 162.0, 164.0 ppm;
HRMS (ESI): calcd for C17H13F2N2O2 [M+H]+: 315.0945, found:
315.0943; MS (ESI): m/z=315 [M+H]+.
Method E: CDI reaction in THF. CDI (2 equiv) was added to a solu-
tion of the corresponding alcohol (1 equiv) in THF (10 mLmmolꢀ1).
The solution was then heated at reflux (708C) for four days. After
cooling to room temperature, the mixture was poured into H2O
and extracted with CH2Cl2 (3ꢁ25 mL). The combined organic
phases were washed with brine, dried over MgSO4 and evaporated
under reduced pressure. Then the desired product was purified by
chromatography on silica gel.
(4’-Fluorobiphenyl-4-yl)(1H-imidazol-5-yl)methanol (23). A solu-
tion of 23a (0.10 g, 0.20 mmol) in MeOH (5 mL) was stirred with
pyridinium·HCl (35 mg, 0.30 mmol) at 608C for 4 h. The reaction
was then quenched by adding saturated NaHCO3(aq) (10 mL); EtOAc
(20 mL) was added, and the phases separated. The aqueous phase
was extracted with EtOAc (3ꢁ20 mL), the combined organic ex-
tracts were dried over Na2SO4, and concentrated under reduced
pressure. The crude product was purified by flash chromatography;
yield: 0.05 g (91%); colorless oil; Rf =0.19 (EtOAc/MeOH, 95:5);
1H NMR (CDCl3+[D6]DMSO, 500 MHz): d=2.99 (s, 1H), 5.84 (s, 1H),
6.87 (s, 2H), 7.02–7.06 (m, 2H), 7.41–7.47 ppm (m, 6H); 13C NMR
(CDCl3+[D6]DMSO, 125 MHz): d=71.2 (COH), 115.8, 116.0, 116.4,
127.2, 127.9, 129.1, 129.4, 137.1, 138.6, 140.7, 142.3, 150.9, 162.1,
164.1 ppm; HRMS (ESI): calcd for C16H14FN2O [M+H]+: 269.1090,
found: 269.1088; MS (ESI):m/z=269 [M+H]+.
Imidazole-1-carboxylic acid 1-(3’,4’-difluorobiphenyl-4-yl)propyl
ester (3). Synthesized according to Method E using 2a (0.10 g,
0.42 mmol) and CDI (0.14 g, 0.83 mmol); yield: 0.09 g (37%); white
1
solid, mp: 49–508C; Rf =0.29 (CH2Cl2/MeOH, 95:5); H NMR (CDCl3,
500 MHz): d=1.01 (t, J=14.8 Hz, 3H, CH3), 2.01–2.21 (q, J=14.8,
15.5 Hz, 2H, CH2), 5.88 (t, J=15.5 Hz, 1H, CH), 7.15 (s, 1H), 7.22–
7.29 (m, 2H), 7.32–7.34 (m, 1H), 7.36–7.38 (m, 1H), 7.46 (d, J=
8.2 Hz, 2H), 7.54 (d, J=8.2 Hz, 2H), 8.35 ppm (s, 1H, Im-2H);
13C NMR (CDCl3, 125 MHz): d=9.9 (CH3), 28.9 (CH2), 82.6 (CH), 115.9,
116.1, 117.5, 117.7, 122.4, 122.9, 123.0, 127.3, 127.4, 129.5, 137.8,
140.2, 149.0, 149.1, 151.2 ppm; HRMS (ESI): calcd for C19H17F2N2O2
[M+H]+: 343.1258, found: 343.1250; MS (ESI): m/z=343 [M+H]+.
Imidazole-1-carboxylic acid 1-(2’,4’-difluorobiphenyl-4-yl)propyl
ester (5). Synthesized according to Method E using 4a (0.21 g,
0.85 mmol) and CDI (0.28 g, 1.69 mmol); yield: 0.11 g (43%); white
5-[1-(4’-Fluorobiphenyl-4-yl)-2-methylprop-1-enyl]-1H-imidazole
(25). Compound 24 (0.10 mg, 0.32 mol) was held at reflux in an
iPrOH solution of HCl (10 mL, 3n) for 2 h. Afterward, the resulting
solution was concentrated under reduced pressure and washed
with Et2O (3ꢁ25 mL). No further purification was necessary; yield:
0.07 g (77%); colorless oil; Rf =0.29 (EtOAc/MeOH, 95:5); 1H NMR
(CDCl3, 500 MHz): d=1.80 (s, 3H), 2.05 (s, 3H), 6.98 (s, 1H), 7.12
(dd, J=8.7, 8.8 Hz, 2H), 7.19 (d, J=8.3 Hz, 2H), 7.50 (d, J=8.3 Hz,
2H), 7.54 (dd, J=5.3, 8.7 Hz, 2H), 7.63 ppm (s, 1H); 13C NMR (CDCl3,
125 MHz): d=22.7 (CH3), 23.1 (CH3), 115.6, 115.8, 123.4, 125.2,
126.9, 128.5, 128.5, 129.8, 130.3, 132.6, 133.6, 133.7, 134.9, 136.6,
136.7, 136.7, 137.1, 138.8, 139.8, 160.9, 161.5 ppm; 19F NMR (CDCl3,
400 MHz): d=ꢀ115.56 ppm (s, 1F); HRMS (ESI): calcd for C19H18FN2
[M+H]+: 293.1454, found: 293.1455; MS (ESI):m/z=293 [M+H]+.
1
solid, mp: 76–778C; Rf =0.29 (CH2Cl2/MeOH, 95:5); H NMR (CDCl3,
500 MHz): d=1.01 (t, J=14.8 Hz, 3H, CH3), 2.03–2.19 (q, J=14.8,
15.5 Hz, 2H, CH2), 5.87 (t, J=15.5 Hz, 1H, CH), 6.89–7.06 (m, 2H),
7.09–7.11 (m, 1H), 7.36–7.40 (m, 1H), 7.45–7.48 (m, 3H), 7.52 (dd,
J=1.6, 1.9 Hz, 2H), 8.24 ppm (s, 1H, Im-2H); 13C NMR (CDCl3,
125 MHz): d=9.9 (CH3), 28.9 (CH2), 82.2 (CH), 104.4, 111.7, 117.2,
124.5, 126.8, 129.3, 130.7, 135.5, 136.9, 137.9, 148.0, 158.5, 1160.7,
161.4, 163.4, 171.2 ppm; HRMS (ESI): calcd for C19H17F2N2O2 [M+H]+
: 343.1258, found: 343.1247; MS (ESI): m/z=343 [M+H]+.
Imidazole-1-carboxylic acid 1-(2’,5’-difluorobiphenyl-4-yl)propyl
ester (8). Synthesized according to Method E using 7a (0.12 g,
0.46 mmol) and CDI (0.15 g, 0.93 mmol); yield: 0.10 g (69%); white
1
solid, mp: 77–788C; Rf =0.29 (CH2Cl2/MeOH, 95:5); H NMR (CDCl3,
5-[1-(4’-Fluorobiphenyl-4-yl)-2-methylpropyl]-1H-imidazole (26).
Pearlman’s catalyst (5 mg, 7.12 mmol) and 25 (50 mg, 0.17 mmol)
were prepared in EtOH and THF (2:1, 5 mL) under H2 atmosphere.
500 MHz): d=1.01 (t, J=14.8 Hz, 3H, CH3), 2.02–2.20 (q, J=14.8,
15.5 Hz, 2H, CH2), 5.88 (t, J=15.5 Hz, 1H, CH), 6.93–7.02 (m, 1H),
908
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910

Fluorine Substitution in CYP17 Inhibitors
The mixture was left stirring for 3 h, then the catalyst was filtered
off three times, and the solution was concentrated under reduced
pressure. The obtained solid was washed with Et2O (3ꢁ25 mL). No
further purification was necessary; yield: 50 mg (100%); yield:
0.07 g (77%); colorless oil; Rf =0.29 (EtOAc/MeOH, 95:5); 1H NMR
(CDCl3, 500 MHz): d=0.86 (d, J=6.6 Hz, 3H, CH3), 0.99 (d, J=
6.6 Hz, 3H, CH3), 3.60 (d, J=9.4 Hz, 1H, CH), 6.90 (s, 1H), 7.09 (dd,
J=8.7, 8.8 Hz, 2H), 7.36 (d, J=8.2 Hz, 2H), 7.45 (d, J=8.2 Hz, 2H),
7.51 (dd, J=5.4, 8.7 Hz, 2H), 7.55 ppm (s, 1H); 13C NMR (CDCl3,
125 MHz): d=21.2, 21.8, 32.5, 51.9, 115.4, 115.6, 123.1, 126.9, 128.4,
128.5, 128.9, 134.4, 137.1, 138.1, 142.4 ppm; HRMS (ESI): calcd for
C19H20FN2 [M+H]+: 295.1611, found: 295.1605; MS (ESI):m/z=295
[M+H]+.
MEP: For each docked compound, geometry optimization was per-
formed by using the B3LYP hybrid functional[20] in combination
with a 6-311++G (d,p) basis set using the Gaussian 03 software
package.[21] The molecular electrostatic potential (MEP) maps were
plotted using GaussView 3.0, the 3D molecular graphics package of
Gaussian. These electrostatic potential surfaces were generated by
mapping 6-311++G** electrostatic potentials onto surfaces of
molecular electron density (isovalue=0.004 electronꢂꢀ1) and the
ESP values on the surface are color coded, ranging from ꢀ6 kcal
molꢀ1 (red) to +12 kcalmolꢀ1 (blue).
Acknowledgements
The authors thank Dr. Ursula Mꢀller-Vieira and Maria-Christina
Scherzberg for IC50 determinations, Dr. Klaus Biemel for perform-
ing the in vivo tests, and Dr. Kerstin Jahn-Hoffmann, Dr. Mariano
E. Pinto-Bazurco Mendieta and Dr. Marc Bartels for help in the
synthesis of some compounds. The authors also appreciate the
kind help from Prof. Dr. Rolf Mꢀller (Saarland University, Saar-
brꢀcken, Germany) in performing HRMS.
Docking studies
Ligands: All molecular modeling studies were performed on an
Intel P4 CPU 3.00 GHz running Linux CentOS5.2. The structures of
the inhibitors were built with SYBYL 8 (Sybyl, Tripos Inc., St. Louis,
MO, USA) and energy minimized in the MMFF94s force field[25] as
implemented in SYBYL.
Docking: Molecular docking calculations were performed for vari-
ous inhibitors listed in Table 1. Because the GOLD docking program
allows flexible docking of the compounds, no conformational
search was employed to the ligand structures. GOLD gave the best
poses by a genetic algorithm (GA) search strategy. Ligands were
docked in 50 independent genetic algorithm runs for each of the
three GOLD docking runs. Heme iron was chosen as the active site
origin, while the radius was at 19 ꢂ. The automatic active site de-
tection was switched on. A distance constraint of a minimum of
1.9 and a maximum of 2.5 ꢂ between the sp2-hybridised nitrogen
of the imidazole and the iron was set. Additionally, the goldscor-
e.p450_pdb parameters were used, and some of the GoldScore pa-
rameters were modified to improve the weight of hydrophobic in-
teraction and of the coordination between iron and nitrogen. The
genetic algorithm default parameters were set as suggested by the
GOLD authors. On the other hand, the annealing parameters of fit-
ness function were set at 3.5 ꢂ for hydrogen bonding and 6.5 ꢂ for
van der Waals interactions.
Keywords: anticancer agents · cytochrome P450 inhibitors ·
fluorine · pharmacokinetics · prostate cancer
[1] A. Jemal, R. Siegel, E. Ward, Y. Hao, J. Xu, T. Murray, M. J. Thun, CA
Cancer J. Clin. 2008, 58, 71–96.
[2] I. Huhtaniemi, H. Nikula, M. Parvinen, S. Rannikko, Am. J. Clin. Oncol.
1988, 11 (Suppl. 1), S11–S15.
[3] F. Labrie, A. Dupont, A. Belanger, L. Cusan, Y. Lacourciere, G. Monfette,
J. G. Laberge, J. P. Emond, A. T. Fazekas, J. P. Raynaud, J. M. Husson, Clin.
Invest. Med. 1982, 5, 267–275.
[4] a) T. Hara, J. Miyazaki, H. Araki, M. Yamaoka, N. Kanzaki, M. Kusaka, M.
Miyamoto, Cancer Res. 2003, 63, 149–153; b) X. Y. Zhao, P. J. Malloy,
A. V. Krishnan, S. Swami, N. M. Navone, D. M. Peehl, D. Feldman, Nat.
Med. 2000, 6, 703–706.
[5] a) C. I. Herold, K. L. Blackwell, Clin. Breast Cancer 2008, 8, 50–64; b) R. W.
Hartmann, H. Bayer, G. Grꢃn, J. Med. Chem. 1994, 37, 1275–1281; c) F.
Leonetti, A. Favia, A. Rao, R. Aliano, A. Paluszcak, R. W. Hartmann, A.
Carotti, J. Med. Chem. 2004, 47, 6792–6803; d) M. Le Borgne, P. March-
and, M. Duflos, B. Delevoye-Seiller, S. Piessard-Robert, G. Le Baut, R. W.
Hartmann, M. Palzer, Arch. Pharm. 1997, 330, 141–145; e) M. P. Lꢄzꢄ, M.
Le Borgne, P. Pinson, A. Palusczak, M. Duflos, G. Le Baut, R. W. Hart-
mann, Bioorg. Med. Chem. Lett. 2006, 16, 1134–1137; f) S. Gobbi, A. Cav-
alli, A. Rampa, F. Belluti, L. Piazzi, A. Paluszcak, R. W. Hartmann, M. Reca-
natini, A. Bisi, J. Med. Chem. 2006, 49, 4777–4780; g) A. Cavalli, A. Bisi,
C. Bertucci, C. Rosini, A. Paluszcak, S. Gobbi, E. Giorgio, A. Rampa, F. Bel-
luti, L. Piazzi, P. Valenti, R. W. Hartmann, M. Recanatini, J. Med. Chem.
2005, 48, 7282–7289; h) S. Castellano, G. Stefancich, R. Ragno, K.
Schewe, M. Santoriello, A. Caroli, R. W. Hartmann, G. Sbardella, Bioorg.
Med. Chem. 2008, 16, 8349–8358.
[6] a) S. Lucas, R. Heim, C. Ries, K. E. Schewe, B. Birk, R. W. Hartmann, J.
Med. Chem. 2008, 51, 8077–8087; b) S. Lucas, R. Heim, M. Negri, I.
Antes, C. Ries, K. E. Schewe, A. Bisi, S. Gobbi, R. W. Hartmann, J. Med.
Chem. 2008, 51, 6138–6149; c) R. Heim, S. Lucas, C. M. Grombein, C.
Ries, K. E. Schewe, M. Negri, U. Mꢃller-Vieira, B. Birk, R. W. Hartmann, J.
Med. Chem. 2008, 51, 5064–5074; d) M. Voets, I. Antes, C. Scherer, U.
Mꢃller-Vieira, K. Biemel, S. Marchais-Oberwinkler, R. W. Hartmann, J.
Med. Chem. 2006, 49, 2222–2231; e) M. Voets, I. Antes, C. Scherer, U.
Mꢃller-Vieira, K. Biemel, C. Barassin, S. Marchais-Oberwinkler, R. W. Hart-
mann, J. Med. Chem. 2005, 48, 6632–6642; f) S. Ulmschneider, U.
Mꢃller-Vieira, C. D. Klein, I. Antes, T. Lengauer, R. W. Hartmann, J. Med.
Chem. 2005, 48, 1563–1575; g) S. Ulmschneider, U. Mꢃller-Vieira, M. Mi-
trenga, R. W. Hartmann, S. Oberwinkler-Marchais, C. D. Klein, M. Bureik,
R. Bernhardt, I. Antes, T. Lengauer, J. Med. Chem. 2005, 48, 1796–1805.
Analogously as for GOLD, no conformational search was performed
prior docking with FlexX, as the ligands are docked according to
an incremental fragment docking strategy. Standard parameter set-
tings were used except for “base placement”, which was set on
single interaction scan, and “chemical parameters”, in which the
maximum overlap volume of the subroutine “clash handling” was
set at a range of 3.6 ꢂ. Additionally the “FlexX-Pharm” module was
employed, setting the heme iron as an octahedral coordinating
metal pharmacophore point. The very same iron atom was chosen
as active site center, and amino acid residues within 16 ꢂ were
considered as part of the active site.
All the poses resulting from three docking runs (GOLD–ChemScore,
GOLD–GoldScore, and FlexX) for each compound were clustered
with ACIAP,[19] and the representative structure of each significant
cluster was selected. After the docking simulations and cluster
analysis were performed, the quality of the docked representative
poses was evaluated based on visual inspection of the putative
binding modes of the ligands. The latter were further evaluated by
MOE (http://www.chemcmpd.com) with its LigX module and evalu-
ated by means of the various scoring functions (GoldScore, Chem-
Score, and the empirical FlexX Score).
ChemMedChem 2010, 5, 899 – 910
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
909

R. W. Hartmann et al.
MED
[7] B. C. Chung, J. Picardo-Leonard, M. Hanui, M. Bienkowski, P. F. Hall, J. E.
Shively, W. L. Miller, Proc. Natl. Acad. Sci. USA 1987, 84, 407–411.
[8] a) K. A. Harris, V. Weinberg, R. A. Bok, M. Kakefuda, E. J. Small, J. Urol.
2002, 168, 542–545; b) J. Eklund, M. Kozloff, J. Vlamakis, A. Starr, M.
Mariott, L. Gallot, B. Jovanovic, L. Schilder, E. Robin, M. Pins, R. C.
Bergan, Cancer 2006, 106, 2459–2465.
[9] a) G. A. Potter, S. E. Banie, M. Jarman, M. G. Rowlands, J. Med. Chem.
1995, 38, 2463–2471; b) J. P. Burkhart, C. A. Gates, M. E. Laughlin, R. J.
Resvick, N. P. Peet, Bioorg. Med. Chem. 1996, 4, 1411–1420; c) V. D. Han-
dratta, T. S. Vasaitis, V. C. O. Njar, L. K. Gediya, R. Kataria, P. Chopra, D.
Newman, Jr., R. Farquhar, Z. Guo, Y. Qiu, A. M. H. Brodie, J. Med. Chem.
2005, 48, 2972–2984; d) V. C. O. Njar, K. Kato, I. P. Nnane, D. N. Grigor-
yev, B. J. Long, A. M. H. Brodie, J. Med. Chem. 1998, 41, 902–912.
[10] a) R. W. Hartmann, M. Hector, M. Haidar, P. B. Ehmer, W. Reichert, J. Jose,
J. Med. Chem. 2000, 43, 4266–4277; b) R. W. Hartmann, M. Hector, B. G.
Wachall, A. Palusczak, M. Palzer, V. Huch, M. Veith, J. Med. Chem. 2000,
43, 4437–4445; c) V. C. O. Njar, M. Hector, R. W. Hartmann, Bioorg. Med.
Chem. 1996, 4, 1447–1453; d) S. Haidar, P. B. Ehmer, R. W. Hartmann,
Arch. Pharm. 2001, 334, 373–374.
M. S. South, J. Med. Chem. 2003, 46, 4297–4312; l) J. J. Parlow, R. G. Kur-
umbail, R. A. Stegeman, A. M. Stevens, W. C. Stallings, M. S. South, J.
Med. Chem. 2003, 46, 4696–4701; m) B. K. Park, N. R. Kitteringham, P. M.
O’Neill, Annu. Rev. Pharmacol. Toxicol. 2001, 41, 443–470.
[14] N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457–2483.
[15] Y. Tang, Y. Dong, J. L. Vennerstrom, Synthesis 2004, 15, 2540–2544.
[16] P. B. Ehmer, J. Jose, R. W. Hartmann, J. Steroid Biochem. Mol. Biol. 2000,
75, 57–63.
[17] a) G. Jones, P. Willett, R. C. Glen, A. R. Leach, R. Taylor, J. Mol. Biol. 1997,
267, 727–748; b) M. L. Verdonk, J. C. Cole, M. J. Hartshorn, C. W. Murray,
R. D. Taylor, Proteins 2003, 52, 609–623.
[18] a) B. Kramer, M. Rarey, T. Lengauer, Proteins Struct. Funct. Genet. 1999,
37, 228–241; b) FlexX trial version 3.1.3, BioSolveIT GmbH, Sankt Augus-
tin (Germany), http://www.biosolveit.de/ (accessed March 7, 2010).
[19] a) G. Bottegoni, A. Cavalli, M. A. Recanatini, J. Chem. Inf. Model. 2006, 46,
852–862; b) G. Bottegoni, W. Rocchia, M. Recanatini, A. Cavalli, Bioinfor-
matics 2006, 22, e58–e65.
[20] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) P. J. Stephens, F. J.
Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys. Chem. 1994, 98, 11623–
11627.
[21] Gaussian 03, Revision C.02: M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsu-
ji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Na-
kajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Och-
terski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui,
A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian Inc., Wallingford,
CT (USA), 2004.
[22] GaussView 3.0: I. Dennington, K. T. Roy, J. Millam, K. Eppinnett, W. L.
Howell, R. Gilliland, Semichem Inc., Shawnee Mission, KS (USA), 2003.
[23] a) E. Bey, S. Marchais-Oberwinkler, M. Negri, P. Kruchten, A. Oster, T.
Klein, A. Spadaro, R. Werth, M. Frotscher, B. Birk, R. W. Hartmann, J. Med.
Chem. 2009, 52, 6724–6743; b) P. Kruchten, R. Werth, E. Bey, A. Oster, S.
Marchais-Oberwinkler, M. Frotscher, R. W. Hartmann, J. Steroid Biochem.
Mol. Biol. 2009, 114, 200–206; c) E. Bey, S. Marchais-Oberwinkler, R.
Werth, M. Negri, Y. A. Al-Soud, P. Kruchten, A. Oster, M. Frotscher, B. Birk,
R. W. Hartmann, J. Med. Chem. 2008, 51, 6725–6739; d) S. Marchais-
Oberwinkler, P. Kruchten, M. Frotscher, E. Ziegler, A. Neugebauer, U.
Bhoga, E. Bey, U. Mꢃller-Vieira, J. Messinger, H. Thole, R. W. Hartmann, J.
Med. Chem. 2008, 51, 4685–4698; e) M. Frotscher, E. Ziegler, S. March-
ais-Oberwinkler, P. Kruchten, A. Neugebauer, L. Fetzer, C. Scherer, U.
Mꢃller-Vieira, J. Messinger, H. Thole, R. W. Hartmann, J. Med. Chem.
2008, 51, 2158–2169.
[11] a) G. A. Wꢅchter, R. W. Hartmann, T. Sergejew, G. L. Grꢃn, D. Ledergerber,
J. Med. Chem. 1996, 39, 834–841; b) F. C. Y. Chan, G. A. Potter, S. E.
Barrie, B. P. Haynes, M. G. Rowlands, G. Houghton, M. Jarman, J. Med.
Chem. 1996, 39, 3319–3323.
[12] a) B. G. Wachall, M. Hector, Y. Zhuang, R. W. Hartmann, Bioorg. Med.
Chem. 1999, 7, 1913–1924; b) Y. Zhuang, B. G. Wachall, R. W. Hartmann,
Bioorg. Med. Chem. 2000, 8, 1245–1252; c) F. Leroux, T. U. Hutschenreut-
er, C. Charriere, R. Scopelliti, R. W. Hartmann, Helv. Chim. Acta 2003, 86,
2671–2686; d) T. U. Hutschenreuter, P. E. Ehmer, R. W. Hartmann, J.
Enzyme Inhib. Med. Chem. 2004, 19, 17–32; e) C. Jagusch, M. Negri, U. E.
Hille, Q. Hu, M. Bartels, K. Jahn-Hoffmann, M. A. E. Pinto-Bazurco Men-
dieta, B. Rodenwaldt, U. Mꢃller-Vieira, D. Schmidt, T. Lauterbach, M. Re-
canatini, A. Cavalli, R. W. Hartmann, Bioorg. Med. Chem. 2008, 16, 1992–
2010; f) M. A. E. Pinto-Bazurco Mendieta, M. Negri, C. Jagusch, U. E.
Hille, U. Mꢃller-Vieira, D. Schmidt, K. Hansen, R. W. Hartmann, Bioorg.
Med. Chem. Lett. 2008, 18, 267–273; g) Q. Hu, M. Negri, K. Jahn-Hoff-
mann, Y. Zhuang, S. Olgen, M. Bartels, U. Mꢃller-Vieira, T. Lauterbach,
R. W. Hartmann, Bioorg. Med. Chem. 2008, 16, 7715–7727; h) M. A. E.
Pinto-Bazurco Mendieta, M. Negri, C. Jagusch, U. Mꢃller-Vieira, T. Lauter-
bach, R. W. Hartmann, J. Med. Chem. 2008, 51, 5009–5018; i) U. E. Hille,
Q. Hu, C. Vock, M. Negri, M. Bartels, U. Mꢃller-Vieira, T. Lauterbach, R. W.
Hartmann, Eur. J. Med. Chem. 2009, 44, 2765–2775; j) M. A. E. Pinto-Ba-
zurco Mendieta, M. Negri, Q. Hu, U. E. Hille, C. Jagusch, K. Jahn-Hoff-
mann, U. Mꢃller-Vieira, D. Schmidt, T. Lauterbach, R. W. Hartmann, Arch.
Pharm. 2008, 341, 597–609; k) U. E. Hille, Q. Hu, M. A. E. Pinto-Bazurco
Mendieta, M. Bartels, C. A. Vock, T. Lauterbach, R. W. Hartmann, Comptes
Rendus Chim. 2009, 12, 1117–1126.
[13] a) W. K. Hagmann, J. Med. Chem. 2008, 51, 4359–4369; b) H. J. Bçhm, D.
Banner, S. Bendels, M. Kansy, B. Kuhn, K. Mꢃller, U. Obst-Sander, M.
Stahl, ChemBioChem 2004, 5, 637–643; c) K. Mꢃller, C. Faeh, F. Dieder-
ich, Science 2007, 317, 1881–1886; d) S. DiMagno, H. Sun, Curr. Top.
Med. Chem. 2006, 6, 1473–1482; e) R. Paulini, K. Mꢃller, F. Diederich,
Angew. Chem. 2005, 117, 1820–1839; Angew. Chem. Int. Ed. 2005, 44,
1788–1805; f) J. A. Olsen, D. W. Banner, P. Seiler, U. O. Sander, A. D’Arcy,
M. Stihle, K. Mꢃller, F. Diederich, Angew. Chem. 2003, 115, 2611–2615;
Angew. Chem. Int. Ed. 2003, 42, 2507–2511; g) J. A. Olsen, D. W. Banner,
P. Seiler, B. Wagner, T. Tschopp, U. Obst-Sander, M. Kansy, K. Mꢃller, F.
Diederich, ChemBioChem 2004, 5, 666–675; h) F. Hof, D. M. Scofield,
W. B. Schweizer, F. Diederich, Angew. Chem. 2004, 116, 5166–5169;
Angew. Chem. Int. Ed. 2004, 43, 5056–5059; i) E. S. Istvan, J. Deisenhofer,
Science 2001, 292, 1160–1164; j) G. Phillips, D. D. Davey, K. A. Eagen,
S. K. Koovakkat, A. Liang, H. P. Ng, M. Pinkerton, L. Trinh, M. Whitlow,
A. M. Beatty, M. M. Morrissey, J. Med. Chem. 1999, 42, 1749–1756; k) J. J.
Parlow, A. M. Stevens, R. A. Stegeman, W. C. Stallings, R. G. Kurumbail,
[24] a) F. Picard, T. Schulz, R. W. Hartmann, Bioorg. Med. Chem. 2002, 10,
437–448; b) E. Baston, A. Palusczak, R. W. Hartmann, Eur. J. Med. Chem.
2000, 35, 931–940; c) W. Reichert, J. Jose, R. W. Hartmann, Arch. Pharm.
2000, 333, 201–204; d) F. Picard, E. Baston, W. Reichert, R. W. Hartmann,
Bioorg. Med. Chem. 2000, 8, 1479–1487; e) E. Baston, R. W. Hartmann,
Bioorg. Med. Chem. Lett. 1999, 9, 1601–1606.
[25] T. A. Halgren, J. Comput. Chem. 1999, 20, 730–748.
Received: February 15, 2010
Revised: March 22, 2010
Published online on April 30, 2010
910
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 899 – 910