Subtle Structural Differences Trigger Inhibitory Activity of Propafenone Analogues at the Two Polyspecific ABC Transporters: P-Glycoprotein (P-gp) and Breast Cancer Resistance Protein (BCRP)

Article abstract of DOI:10.1002/cmdc.201500592

The transmembrane ABC transporters P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP) are widely recognized for their role in cancer multidrug resistance and absorption and distribution of compounds. Furthermore, they are linked to drug–drug interactions and toxicity. Nevertheless, due to their polyspecificity, a molecular understanding of the ligand-transporter interaction, which allows designing of both selective and dual inhibitors, is still in its infancy. This study comprises a combined approach of synthesis, in silico prediction, and in vitro testing to identify molecular features triggering transporter selectivity. Synthesis and testing of a series of 15 propafenone analogues with varied rigidity and basicity of substituents provide first trends for selective and dual inhibitors. Results indicate that both the flexibility of the substituent at the nitrogen atom, as well as the basicity of the nitrogen atom, trigger transporter selectivity. Furthermore, inhibitory activity of compounds at P-gp seems to be much more influenced by logP than those at BCRP. Exploiting these differences further should thus allow designing specific inhibitors for these two polyspecific ABC-transporters.

Full text of DOI:10.1002/cmdc.201500592

DOI: 10.1002/cmdc.201500592
Full Papers
Subtle Structural Differences Trigger Inhibitory Activity of
Propafenone Analogues at the Two Polyspecific ABC
Transporters: P-Glycoprotein (P-gp) and Breast Cancer
Resistance Protein (BCRP)
Theresa Schwarz,[a] Floriane Montanari,[a] Anna Cseke,[a] Katrin Wlcek,[a] Lene Visvader,[a]
Sarah Palme,[a] Peter Chiba,[b] Karl Kuchler,[c] Ernst Urban,[a] and Gerhard F. Ecker*[a]
The transmembrane ABC transporters P-glycoprotein (P-gp)
and breast cancer resistance protein (BCRP) are widely recog-
nized for their role in cancer multidrug resistance and absorp-
tion and distribution of compounds. Furthermore, they are
linked to drug–drug interactions and toxicity. Nevertheless,
due to their polyspecificity, a molecular understanding of the
ligand-transporter interaction, which allows designing of both
selective and dual inhibitors, is still in its infancy. This study
comprises a combined approach of synthesis, in silico predic-
tion, and in vitro testing to identify molecular features trigger-
ing transporter selectivity. Synthesis and testing of a series of
15 propafenone analogues with varied rigidity and basicity of
substituents provide first trends for selective and dual inhibi-
tors. Results indicate that both the flexibility of the substituent
at the nitrogen atom, as well as the basicity of the nitrogen
atom, trigger transporter selectivity. Furthermore, inhibitory ac-
tivity of compounds at P-gp seems to be much more influ-
enced by logP than those at BCRP. Exploiting these differences
further should thus allow designing specific inhibitors for these
two polyspecific ABC-transporters.
Introduction
P-glycoprotein (P-gp, ABCB1) and the breast cancer resistance
protein (BCRP, ABCG2) belong to the superfamily of ABC pro-
teins. In humans, this superfamily is composed of 49 members,
subdivided into 7 subfamilies named ABCA to ABCG.[1] Based
on their structural organization, ABC transporters form a passa-
geway across cellular membranes enabling the transport of
substrates. They consist of two transmembrane (TMDs) and
two nucleotide binding domains (NBDs), whereby ATP hydroly-
sis at the NBDs provides the energy necessary for the transport
process.[1]
P-gp and BCRP are efflux transporters, mediating transport
of their substrates out of cells in which they are expressed.
Thus, they are physiologically expressed in organs with barrier
and/or excretion functions including the gastrointestinal (GI)
tract, the blood-brain barrier, the liver, and the kidney.[2]
In the GI tract, both ABC family members are located in the
apical membrane of enterocytes, facing towards the intestinal
lumen, affecting the absorption of compounds from the intes-
tinal lumen into the blood. Similarly, P-gp and BCRP are local-
ized at the apical side of proximal tubule cells and hepatocytes
in the kidney and the liver, respectively. There, they are in-
volved in the excretion of endogenous and exogenous com-
pounds, as well as of metabolites, into the urine or the bile.[3]
At the blood–brain barrier, P-gp and BCRP are located at the
apical/luminal side in the endothelial cells of brain capillaries,
where they protect the brain from xenobiotics.[3]
[a] T. Schwarz, F. Montanari, A. Cseke, Dr. K. Wlcek, L. Visvader, S. Palme,
Prof. E. Urban, Prof. G. F. Ecker
Department of Pharmaceutical Chemistry
Faculty of Life Sciences, University of Vienna
Althanstraße 14, 1090 Vienna (Austria)
E-mail: gerhard.f.ecker@univie.ac.at
[b] Prof. P. Chiba
Department of Medicinal Chemistry, Medical University Vienna
Währingerstraße 10, 1090 Vienna (Austria)
[c] Prof. K. Kuchler
Based on their physiological expression, P-gp and BCRP are
important determinants in absorption, distribution, and excre-
tion of drugs and metabolites that are P-gp and/or BCRP sub-
strates. Inhibition of the proper transport function of these
ABC proteins has been reported to be responsible for drug–
drug interactions.[4] In addition, P-gp and BCRP are also known
for their involvement in multidrug resistance (MDR).[5] As efflux
transporters for different anticancer agents, including anthracy-
clines, kinase inhibitors (imatinib), and alkaloids (e.g. topote-
can),[6] P-gp and/or BCRP overexpression increases the trans-
port of these agents out of the tumor cells resulting in de-
creased drug response.[5]
Department of Medical Biochemistry, Max F. Perutz Laboratories
Medical University Vienna
Dr. Bohr-Gasse 9/2, 1030 Vienna (Austria)
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cmdc.201500592.
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in
any medium, provided the original work is properly cited.
This article is part of a Special Issue on Polypharmacology and
Multitarget Drugs. To view the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cmdc.v11.12/issuetoc.
ChemMedChem 2016, 11, 1380 – 1394
1380 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
With the idea to tackle MDR, several medicinal chemistry ef-
forts have already been taken to create compounds that
would specifically inhibit either P-gp or BCRP individually, or in
combination. Brooks and colleagues[7] designed and tested 18
taxane derivatives for their inhibition of P-gp, MRP1, and BCRP.
They found that taxane derivatives are substrates for P-gp, and
to a lesser extent for MRP1, explaining their MDR reversal
effect by competitive binding. With respect to BCRP, however,
the inhibition was found to be noncompetitive, and the au-
thors suggested that taxane derivatives only bind to one bind-
ing site on BCRP and are not transported. Colabufo and col-
leagues[8] were looking for selective P-gp inhibitors since they
expected that the in vivo toxicity of existing MDR modulating
agents is related to the inhibition of other ABC-transporters
that are instrumental in protecting noncancerous cells. They
synthesized a series of arylmethoxy and arylmethylaminederi-
vatives that showed good P-gp inhibition. Specificity for P-gp
as compared with BCRP seemed to be conferred by the linker,
that is, only arylmethylamine derivatives showed activity for
BCRP. Sim and colleagues[9] focused on a type of flavonoids,
aurones, which are known inhibitors of BCRP and tried to
probe into the selectivity profile towards P-gp by generating
more than one hundred derivatives. While most compounds
preferentially inhibited BCRP, this publication found methox-
yaurones, methoxyindanones, and methoxyflavones to be
active on both P-gp and BCRP. Kühnle et al[10] found that by
modifying the P-gp inhibitor tariquidar, they could greatly in-
crease the selectivity for BCRP. Valdameri et al synthesized
Figure 1. P-gp and BCRP inhibitors.
structure of fumitremorgin C and its derivative Ko143, which
are selective BCRP inhibitors (Figure 1).
In silico classification models for P-gp and BCRP
In order to assess the likelihood of success, in silico classifica-
tion models for P-gp and BCRP were established, and the in-
hibitor profiles of the compounds of interest were predicted.
Two binary classification models for BCRP and P-gp inhibition
were built on previously published large datasets.[14] Both
models are based on ECFP-like (extended connectivity finger-
print) 1024-bit fingerprints as descriptors and use logistic re-
gression (for BCRP) and support vector machine (for P-gp) as
base classifiers. The models are classification models, thus pre-
dicting whether a compound is likely to be an inhibitor of any
of the two transporters. As threshold for active/inactive an IC50
value of 10 mm was used. The predictive capability of the two
models was evaluated by classical 10-fold cross-validation, re-
sults of which are shown in Table 1.
a series of chalcone derivatives[11] and chromone derivatives[12
]
aiming at selective BCRP inhibitors. In the case of the chal-
cones, they showed the influence of the position of methoxy
substitutions on both activity and cytotoxicity. While investi-
gating the chromone derivatives, they found a very selective
and active derivative for BCRP with a high therapeutic index.
Also in case of propafenones, both compounds showing P-gp
and BCRP specificity could be identified.[13] Here, we extend
these studies and aim to understand the molecular features
triggering transporter selectivity. For this, we combined synthe-
sis of novel propafenone derivatives, in silico prediction
models to prioritize our compounds, and in vitro activity meas-
urements.
The two models were used to subsequently predict the in-
hibition against BCRP and P-gp for a series of compounds
(Figure 2). As outlined above, transporter selectivity for propa-
fenones might be triggered by the substituents in the vicinity
Results and Discussion
Both P-glycoprotein and BCRP show polyspecific ligand recog-
nition patterns with considerable overlap in their substrate
and inhibitor profile. Previous results obtained with propafe-
none-type inhibitors pointed towards the vicinity of the nitro-
gen atom as potential structural moiety for selectivity profiling.
In particular, compounds that contain a nonionizable nitrogen
atom showed high selectivity indices. Furthermore, also the
number of rotatable bonds as well as the number of H-bond
acceptors seemed to influence both activity and selectivity.[13]
Thus, the compound design focused on two main features, the
basicity of the nitrogen atom as well as the rigidity of the sub-
stituents in this region. This was also inspired by the polycyclic
Table 1. Evaluation of the predictive power of the in silico models by 10-
fold cross-validation.[a]
Accuracy
Sensitivity
Specificity
MCC
AUC
BCRP inhibition
P-gp inhibition
0.83
0.87
0.77
0.86
0.87
0.89
0.65
0.75
0.90
0.94
[a] Accuracy: the amount of correct predictions among all predictions;
sensitivity: the proportion of inhibitors correctly identified; specificity: the
proportion of noninhibitors correctly identified; Matthews correlation co-
efficient (MCC): correlation coefficient between the real and predicted
classifications (between À1 and 1); area under the receiver operating
characteristic (ROC) curve (AUC): ranking capability of the model.
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1381 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
whole set of compounds indicate that both dual inhibitors as
well as selective P-gp inhibitors should be present in this com-
pound series.
Synthesis
The target compounds 5, 6, 10, and 11 (Figure 2) were synthe-
sized following two general strategies.
Respective hydroxyphenones 1a–c were prepared from 1-(2-
hydroxy-4-methoxyphenyl)ethanone by condensation with
benzaldehyde and reduction of the double bond by hydroge-
nation.[15] For the synthesis of the aryloxypropanolamines 4
and 6, the ketone functionality in differently substituted 2’-hy-
droxyacetophenones 1a–c was protected as a ketal using eth-
ylene glycol and trimethylorthoformate similar to Ref. [16] as
shown in Scheme 1. Subsequently, the phenols 2a–c were O-
alkylated with epichlorohydrine, and the epoxides 3a–c were
reacted with ammonia in methanol to yield the aryloxypropa-
nolamines 4a–c. Attempts to avoid the step of the protection
of the ketone group failed and led to complex mixtures after
the epoxide ring opening with ammonia.
Figure 2. Synthesized target structures.
Amines 4a–c were converted to the respective amides 5a–
c with N-(tert-butoxycarbonyl)-l-proline (boc-proline) using di-
cyclohexylcarbodiimide (DCC) and 4-dimethylaminopyridine
(DMAP) (see Scheme 2). Unfortunately, the NMR analysis of the
proline derivatives 5a–c proved to be difficult due to the oc-
currence of rotamers that lead to broad signals. In addition,
a mixture of diastereomers was formed that could not be sepa-
rated by column chromatography.
of the nitrogen atom. Flexibility of the substituents was trig-
gered by the introduction of a proline (compounds 5 and 10)
and a diketopiperazine moiety (compounds 6 and 11), where-
by the latter contained only nonbasic nitrogen atoms. Further-
more, also lipophilicity of the compounds was varied by differ-
ent substituents at the phenone moiety (R1). Finally, the linker
region between the nitrogen atom and the central aromatic
ring (compound 5 vs. 10 and 6 vs. 11), as well as the substitu-
tion on the central aromatic ring, was modified. In order to
specifically analyze the influence of the basicity of the nitrogen
atom on the transporter interaction profile of propafenones,
the respective isopropyl, fluoroisopropyl, and difluoroisopropyl
analogues 16a–c were synthesized. The predictions for the
For synthesis of the diketopiperazines 6a–c, amines 4a–
c were reacted with methyl(2-chloroacetyl)prolinate, which was
prepared from proline methyl ester and 2-chloroacetyl chlo-
ride.[17] The diketopiperazine analogues showed NMR spectra
that were easier to interpret due to the fixed conformation.
The last step in both synthetic routes was the deprotection of
Scheme 1. Synthesis of intermediate products 4a–c. Reagents and conditions: a) ethylene glycol, trifluoromethanesulfonic acid (TfOH), trimethylorthoformate,
hexane, 408C, 3 d, 63–99%; b) epichlorohydrine, NaH, DMF, 708C, o/n, 74–99%; c) NH3 in MeOH, RT, 2 d 14–54%.
Scheme 2. Synthetic route for final compounds 5a–c and 6a–c. Reagents and conditions: a) boc-proline, DCC, DMAP, CH2Cl2, 2 h to o/n, 41–71%; b) 4n HCl, di-
oxane, RT, 4 h, 98%–quant.; c) methyl (2-chloroacetyl)prolinate, Et3N, RT, CH3CN, 2.5 d or reflux, 2-ethoxyethanol, 2–3 d, then 2n HCl, CH3CN, RT, 4 h, 6–61% (2
steps).
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1382 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the carbonyl group under acidic conditions, and, for the pro-
line derivatives, formation of a hydrochloride. In some cases,
the deprotection of the ketone occurred already partially
during the column chromatography step.
sicity and thus reactivity of the fluorinated amines. Therefore
the amine 4a was reacted with the respective fluorosubstitut-
ed acetone derivatives in a reductive amination reaction fol-
lowed by deprotection of the ketone (Scheme 5).
For the synthesis of the aryloxyethanolamines 10 and 11,
the substituted 2’-hydroxyphenones 1a–c were first reacted
with bromoacetonitrile, followed by a protection of the car-
bonyl group as outlined above (see Scheme 3). Subsequently,
the nitriles 8a–c were reduced using lithium aluminium hy-
dride[18] to yield the amines 9a–c. Attempts to improve the
yield by using other reducing agents, such as Red-Al (sodium
bis(2-methoxyethoxy)aluminumhydride, used for the synthesis
of 9a) or borane-tetrahydrofuran (THF)-complex failed.
An alternative route for the synthesis of intermediate prod-
ucts 9 comprised the O-alkylation of the phenol 1a with 1,2-di-
bromoethane[19] as shown in Scheme 4. Then the carbonyl
group of 12a is protected with ethylene glycol, and the bro-
mide is substituted with trifluoroacetamide[20] to give 14a. This
derivative can easily be cleaved with base, furnishing the
amine 9a in higher yields and with less tedious steps. The pro-
line amide and diketopiperazine derivatives were synthesized
in analogy to the procedures as described above, and after de-
protection and, if possible, hydrochloride formation, the target
compounds 10a–c and 11 a–c were obtained.
In vitro studies
P-gp and BCRP inhibitory activity was assessed by implement-
ing an intracellular accumulation assay of daunorubicin or mi-
toxantrone in cells overexpressing P-gp or BCRP, respectively.
In a first run, the effect of compounds 5, 6, 10, 11, and 16
(Figure 2) at 100 mm final concentration was tested (Figure 3).
All compounds showing an inhibitory effect of more than 50%
compared with full inhibition by the known inhibitors of P-gp
(verapamil) and BCRP (Ko143) were further processed, and IC50
values were determined (IC50 curves and Hill coefficients are
available in Figure S1 and Table S3 in the Supporting Informa-
tion). In order to allow comparison with the in silico classifica-
tion models, compounds with IC50 values below 10 mm were
considered as active. As shown in Table 2, eight compounds
were identified as inhibitors for BCRP (5b, 6a, 6b, 10a, 10b,
11 b, 16b, 16c), while one compound (5a) showed borderline
activity on BCRP with an IC50 value of 16.13Æ3 mm. For P-gp,
nine compounds were classified as inhibitors (5a, 5b, 6a, 10a,
10b, 11 b, 16a, 16b, 16c). In addition, compounds 6b and
11 a showed borderline inhibitory activity on P-gp with IC50
values of 11 Æ1 and 13.9Æ1.5 mm, respectively.
Following the standard synthetic route for propafenone de-
rivatives, compound 16a was prepared by reacting the epox-
ide 15 with isopropylamine. For the fluorinated derivatives
16b and 16c, this was not possible due to the decreased ba-
Scheme 3. Synthesis of compounds 10a–c and 11 a–c. Reagents and conditions: a) bromoacetonitrile, K2CO3, acetone, 508C, 4 d (88–100%); b) ethylene glycol,
p-toluenesulfonic acid (pTosOH), toluene, reflux, o/n, 72–97%; c) LiAlH4 or Red-Al, Et2O, reflux or RT, 3–4 h, 15-42%; d) boc-proline, DCC, DMAP, CH2Cl2, 2 h to
o/n, 57-78%; e) 4n HCl, dioxane, RT, 4 h, 97%–quant.; f) methyl (2-chloroacetyl)prolinate, Et3N, reflux, 2-ethoxyethanol, 2 d, then 2n HCl, CH3CN, RT, 4 h, 19–
39% (2 steps).
Scheme 4. Alternative route to compound 9a. Reagents and conditions: a) 1,2-dibromoethane, K2CO3, Bu4NBr, toluene, water, reflux, 6 d, 79%; b) ethylene
glycol, TfOH, trimethylorthoformate, hexane, 408C, 3 d, 99%; c) trifluoroacetamide, K2CO3, Bu4NBr, DMF, 808C, 4.5 h, 65%; d) KOH, MeOH, water, RT, 2 h, 99%.
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1383 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 2. Results of in silico prediction and in vitro IC50 evaluation of BCRP
and P-gp inhibition for the synthesized compounds (Figure 2).
Compd logP[a] Substituents
BCRP
P-gp
R1
R2
in silico[b] IC50 [mm][c] in silico[b] IC50 [mm][c]
5a
5b
5c
6a
6b
6c
10a
10b
10c
11 a
11 b
11 c
16a
16b
16c
2.37
2.36
0.53
1.93
1.92
0.09
2.94
2.93
1.10
2.50
2.49
0.66
2.86
3.08
3.47
Bn
Bn
H
Bn
Bn
H
Bn
Bn
H
Bn
Bn
H
H
OMe
H
H
OMe
H
H
OMe
H
H
OMe
H
0.25
0.37
0.08
0.39
0.64
0.15
0.17
0.39
0.07
0.40
0.75
0.19
0.22
0.19
0.24
16.1Æ3.0
3.1Æ0.5
n.d.
5.8Æ1.0
3.8Æ0.6
n.d.
3.3Æ0.6
2.9Æ0.5
n.d.
21.5Æ2.9
2.3Æ0.0
n.d.
0.97
0.92
0.46
0.98
0.96
0.78
0.88
0.83
0.19
0.94
0.91
0.42
0.97
0.97
0.97
2.1Æ0.4
2.2Æ0.6
58.7Æ9.2
1.6Æ0.6
11.0Æ1.0
n.d.
1.3Æ0.1
1.0Æ0.2
34.8Æ1.8
13.9Æ1.5
6.2Æ0.4
n.d.
CH3
CH2F
CH3F
–
–
–
n.d.
0.3Æ0.0
0.5Æ0.1
0.9Æ0.2
5.5Æ0.9
4.9Æ1.1
[a] Calculated partition coefficient logPoctanol/water determined using Molec-
ular Operating Environment (MOE) version 2014.09. [b] Score between 0
and 1, given by in silico models, roughly corresponds to the probability
of being an inhibitor; in silico predictions over 0.5 are marked in bold
and represent compounds that are predicted as inhibitors by the models.
[c] IC50 values represent the meanÆSD of at least n=3 independent ex-
periments performed at least in duplicate; for inhibition of BCRP, 11 a was
measured in n=2 independent experiments; the exact number of experi-
ments is given in the Supporting Information (Figure S1 and Table S3);
IC50 values were evaluated as described in the Experimental Section; n.d.:
not determined.
Scheme 5. Synthesis of compounds 16a–c. Reagents and conditions:
a) iPrNH2, reflux, 21 h, 57%; b) NaBH3CN, mono or difluoroacetone, AcOH,
MeOH, reflux, 4 h, quant.; c) 2n HCl, EtOAc, 2 h, 84–86%.
moiety were the only ones showing a higher inhibitory activity
on BCRP than on P-gp, with 2.9- and 2.7-fold lower IC50 values
for BCRP, respectively. Finally, 5c, 6c, 10c, and 11 c were identi-
fied as noninhibitors for both ABC proteins, as their IC50 values
were either far above 10 mm (5c and 10c on P-gp) or did not
show more than 50% inhibition at 100 mm.
In addition, compound 11 a was shown to be a weak inhibi-
tor for BCRP (IC50: 21.5Æ2.9 mm) and a borderline inhibitor for
P-gp (IC50: 13.9Æ1.5 mm).
Figure 3. Evaluation of compounds used for IC50 measurements. In a first set
of experiments, the effect of all compounds at 100 mm final concentration
was tested on P-gp (dark grey bars) and BCRP (light grey bars) using intra-
cellular accumulation assay as described in the Experimental Section. Data
are shown as % inhibition compared with the positive control verapamil
(P-gp) or Ko143 (BCRP) and are given as meanÆSD of a single experiment
with at least duplicate measurements.
Structure–activity relationships (SAR)
In order to gain insights into molecular features that might
trigger P-gp vs. BCRP selectivity of propafenone-type inhibi-
tors, we modified the substituent at the nitrogen atom to-
wards more rigid and less basic properties. These modifications
were complemented by a set of distinct variations, which
should allow assessing if the compounds still show the basic
SAR pattern reported for propafenones. Specifically, these in-
clude the linker region (propanol vs. ethylene), the phenone
moiety (phenylpropiophenone vs. propiophenone), as well as
the central aromatic ring (4-methoxy vs. 4-H). In addition, we
calculated the logP values of all compounds and put them in
relation to previously published logP/pIC50 correlations for this
scaffold.
Comparing the inhibitory activity on P-gp and BCRP, seven
compounds (5a, 6a, 10a, 10b, 16a, 16b, 16c) were more ef-
fective on P-gp than on BCRP (Table 2). The most pronounced
difference in the inhibitory activity on P-gp and BCRP was ob-
served for the fluorinated compounds 16a, 16b, and 16c and
the aryloxyethanolamine 5a giving more than fivefold higher
IC50 values for BCRP than for P-gp. Another three compounds,
namely 6a, 10a, and 10b were more than 2.6-fold better in-
hibitors for P-gp than for BCRP. Compounds 5b and 11 a
showed comparable inhibitory activity on both ABC proteins
(1.4- and 1.5-fold higher IC50 values for BCRP). The methoxy-
substituted compounds 6b and 11 b bearing a piperazinedione
For P-glycoprotein, the SAR pattern retrieved is mainly in
line with those reported for other propafenone analogues:
i) there is a significant correlation of pIC50 values with the logP
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1384 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
of the compounds (r2 =0.71),[21] ii) elimination of the phenyl
ring of the phenylpropiophenone moiety strongly decreases
biological activity (which is most probably related to the logP
decrease), iii) shortening the distance between the central aro-
matic ring and the nitrogen atom slightly decreases biological
activity,[22] and iv) increasing electron density of the central aro-
matic ring slightly increases activity.[23] However, the ethylene
compound 6a with a diketopiperazine moiety at the nitrogen
atom does not follow this pattern. It is tenfold more active
than its propanol analogue 11 a, and is also more active than
its methoxy analogue 6b by a factor of 7
For BCRP, the picture is more complex, and there are no ob-
vious trends. Also with respect to lipophilicity of the com-
pounds, there is no correlation with pIC50 values (r2 =0.02). In-
terestingly, the most active compound (11 b) belongs to the di-
ketopiperazine series, whereas in case of P-gp it is a proline
analogue (10b).
gp. Thus, these models are still valuable tools for prioritizing
propafenone-type compounds in medicinal chemistry projects.
Selectivity vs. multitarget inhibition
Although the activity differences for compounds 5–11 be-
tween P-gp and BCRP are less than a factor of 10 [selectivity
indices (SI) P-gp/BCRP=0.13–2.82], some trends can be de-
duced from the data. Highest selectivity for P-gp was obtained
for the two ethylene analogues 5a (SI=0.13) and 6a (SI=
0.27), both being unsubstituted at the central aromatic ring
and having either a proline or a diketopiperazine in the vicinity
of the nitrogen atom. In case of BCRP selectivity, the two dike-
topiperazines 6b and 11 b showed the highest SI values (2.82
and 2.70, respectively). Both of them show a methoxy moiety
at position 4 of the central aromatic ring and belong either to
the ethylene (6b) or the propanolamine (11 b) class of com-
pounds. The remaining compounds may be regarded as dual
transporter inhibitors, showing SI values in the range of 0.34–
0.71 and pIC50 values in the low micromolar range. Selectivity
towards BCRP thus seems to be triggered mainly by the sub-
stituent on the central aromatic ring, as well as the diketopi-
perazine moiety. This might at least in part be due to the basic
feature of the nitrogen atom, as exemplified by the 16a–
c series of compounds. By introduction of fluorine atoms, the
basicity of the nitrogen atom is systematically varied. Although
in all cases the compounds are more active with respect to
P-gp than to BCRP, there is a clear tendency in P-gp that lower
basicity/H-bond donor properties lead to lower activity. This is
not seen in BCRP, which is in line with previous observations
for propafenones.[13,24] Furthermore, the switch from proline to
diketopiperazine also decreases flexibility in the vicinity of the
nitrogen atom. It is tempting to speculate that a further
morphing towards a more fumitremorgin C like scaffold would
increase selectivity.
Prospective performance of the in silico models
The BCRP inhibition model tends to give low scores in our
chemical series, that is, rarely predicts a compound as an inhib-
itor. In our specific case, 6 compounds out of 15 are mispre-
dicted as inactives (5b, 6a, 10a, b, 16b, c), which results in an
accuracy of 60%. The P-gp inhibition model, on the contrary,
tends to give high scores, that is, it often predicts a compound
as an inhibitor. In our specific case, the model makes one clear
mistake (predicting the des-phenyl analogue 6c as active) and
2 misclassifications of borderline compounds (predicting 6b
(IC50 =11.0Æ1 mm) and 11 a (IC50 =13.9Æ1.5 mm) as inhibitors
when they show activity slightly above 10 mm, the IC50 thresh-
old applied for categorizing the compounds).
To understand this behavior of both our models, for each
compound of our series we retrieved the five nearest structural
neighbors of the training sets (see Supporting Information for
Tables S1 and S2 showing the average similarity values). In the
BCRP training set, all neighbors retrieved were propafenone
derivatives from the study of Cramer et al.,[13] and all these
compounds were inactives for BCRP. This explains the tenden-
cy of the model to predict most of the propafenone deriva-
tives as noninhibitors of BCRP. In contrast, in the P-gp training
set, all retrieved neighbors were propafenone derivatives anno-
tated as inhibitors of P-gp. This explains the tendency of the
P-gp model to predict most propafenone derivatives as active.
The difference between the prospective results and the pre-
dicted cross-validation results seems striking, but is inherently
linked to the composition of the training set. Indeed, the in sili-
co models are applied on structurally very similar compounds.
As a result, the fingerprint features encoding these structures
are very similar to each other, and the results given by the
models cannot show a high variability. The cross-validation
rather evaluates the generalizability of the model than its spe-
cific capability to properly predict SAR subtleties within the
propafenone family. However, if one directly uses the score to
rank the compounds according to their probability to inhibit
P-gp or BCRP, instead of applying a fixed threshold, the calcu-
lated prospective AUCs would be 0.78 for BCRP and 0.85 for P-
Conclusions
The ABC-transporters P-glycoprotein (P-gp) and the breast
cancer resistance protein (BCRP) are characterized by a broad
and partly overlapping ligand profile. Due to their importance
for absorption, distribution, metabolism, excretion (ADME), and
toxicity, design of multitransporter inhibitors as well as trans-
porter-specific compounds would be a valuable effort for ob-
taining a set of tool compounds. Building on previous work,
we extended our studies on the propafenone scaffold as a ver-
satile chemotype for triggering transporter selectivity and for
increasing our understanding of the molecular features driving
ligand–transporter interaction. Results indicate that both the
flexibility of the substituent at the nitrogen atom, as well as
the basicity of the nitrogen atom, triggers transporter selectivi-
ty. Furthermore, inhibitory activity of compounds for P-gp
seems to be much more influenced by logP than that for
BCRP. Exploiting these differences should thus further allow to
dissect the molecular features triggering ligand selectivity for
the two polyspecific ABC-transporters: P-gp and BCRP.
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1385 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
2-(2-(Oxiran-2-ylmethoxy)phenyl)-2-phenethyl-1,3-dioxolane
(2a): The ketal 1a (5 g, 1 eq, 18.5 mmol) was dissolved in anhy-
drous dimethylformamide (40 mL) and heated to 708C, then NaH
(60% dispersion, 1.11 g, 1.5 eq, 27.7 mmol) was added in portions
during 30 min. After stirring for 1 h at this temperature, epichloro-
hydrine (4.4 mL, 3 eq, 55.5 mmol) was added and heated over-
night. TLC showed no starting material; therefore, the reaction
mixture was concentrated, diluted with Et2O (100 mL) and extract-
ed once with sat. NaHCO3 solution (100 mL) and water (100 mL).
The combined aqueous extracts were extracted once with Et2O
(100 mL), and the combined organic phases were dried over
Na2SO4, and the solvent was evaporated. The crude product was
triturated with petrol ether (20 mL) under cooling, and the petrol
ether phase was removed, which yielded the product as a yellowish
oil (6.0 g, 99%) after drying. NMR data were in good accordance
with the literature.[25]
Experimental Section
Chemistry
General. All moisture-sensitive reactions were conducted in anhy-
drous solvents (Sigma–Aldrich) under dry argon atmosphere. All
solvents and reagents were obtained from Sigma–Aldrich, ABCR,
TCI, or Acros and used without further purification. Reactions were
monitored by analytical thin-layer chromatography (TLC) using pre-
coated TLC sheets ALUGRAM Xtra SIL G/UV254 (Macherey–Nagel,
0.20 mm silica gel 60 layer with fluorescent indicator UV254). Visu-
alization was carried out by UV light (254 nm), ninhydrin, and/or
Seebach stain. Flash chromatography was performed using Merck
silica gel 60m (0.040–0.063 mm). Yields are not optimized.
NMR spectroscopy. NMR spectra were registered either on
a Bruker Avance 200 or Ultrashield 500 spectrometer (Billerica, MA,
USA) at 200 or 500 MHz, using CDCl3 or D6]DMSO as solvents.
Chemical shifts (d) are expressed in parts per million (ppm) relative
to tetramethylsilane (TMS), and coupling constants (J) are given in
Hertz (Hz). Multiplicities are described as s (singlet), brs (broad sin-
glet), d (doublet), t (triplet), q (quadruplet), quin (quintuplet),
2-(4-Methoxy-2-(oxiran-2-ylmethoxy)phenyl)-2-phenethyl-1,3-di-
oxolane (2b): Preparation like 2a; purification by column chroma-
tography (petrol ether:EtOAc, 4:1 to 2:1) afforded a colorless oil
(3.52 g, 74%); 1H NMR (500 MHz, CDCl3): d=2.39–2.49 (m, 2H),
2.59–2.67 (m, 2H), 2.87 (d, J=3.5 Hz, 2H), 3.35–3.39 (m, 1H), 3.80
(s, 3H), 3.85–3.92 (m, 2H), 4.02–4.08 (m, 3H), 4.24 (dd, J=11.2,
3.0 Hz, 1H), 6.47 (dd, J=8.4, 2.4 Hz, 1H), 6.49 (d, J=2.2 Hz, 1H),
7.13 (t, J=7.1 Hz, 1H), 7.17 (d, J=7.3 Hz, 2H), 7.23 (d, J=7.6 Hz,
2H), 7.24–7.29 (m, 1H), 7.42 ppm (d, J=8.5 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=30.2, 39.7, 44.7, 50.2, 55.3, 64.70, 64.72, 68.8,
101.1, 104.2, 109.9, 122.5, 125.4, 128.1 (2C), 128.30, 128.34 (2C),
128.36, 142.4, 157.0, 160.7 ppm; HRMS (ESI): calcd for C21H24NaO5
m/z 357.1697 [M+Na]+, found m/z 357.1697 [M+Na]+.
1
septet, or m (multiplet). H and 13C spectra of tested compounds
are given in the Supporting Information. Purity of target com-
pounds was defined by 13C NMR (absence of additional signals).
High-resolution electrospray ionization mass spectrometry (HR-
ESI-MS). HR-ESI-MS spectra were obtained on a maXis HD ESI-Qq-
TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) using
direct infusion. The ESI ion source was operated as follows: capilla-
ry voltage: 1.0 to 4.0 kV (individually optimized), nebulizer: 0.4 bar
(N2), dry gas flow: 4 LminÀ1 (N2), and dry temperature: 2008C.
Mass spectra were recorded in the range of m/z 50–1550 in the
positive-ion mode. The sum formulas were determined using
Bruker Compass Data Analysis 4.2 based on the mass accuracy
(Dm/zꢀ2 ppm) and isotopic pattern matching (SmartFormula al-
gorithm).
2-Methyl-2-(2-(oxiran-2-ylmethoxy)phenyl)-1,3-dioxolane
(2c):
Preparation like 2a.The crude product was purified by column
chromatography (petrol ether: EtOAc 6:1 to 2:1) to yield 2c as
a colorless oil (3.26 g, 84%);1H NMR (500 MHz, CDCl3): d=1.79 (s,
3H), 2.87–2.92 (m, 2H), 3.38–3.43 (m, 1H), 3.79–3.88 (m, 2H), 4.01–
4.06 (m, 2H), 4.08 (dd, J=11.2, 4.9 Hz, 1H), 4.31 (dd, J=11.2,
3.0 Hz, 1H), 6.92 (d, J=8.2 Hz, 1H), 6.94 (td, J=7.4, 1.1 Hz, 1H),
7.26 (td, J=7.4, 1.1 Hz, 1H), 7.50 ppm (dd, J=7.7, 1.7 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=25.4, 44.7, 50.3, 64.5 (2C), 68.9,
108.5, 113.6, 120.8, 126.8, 129.4, 131.1, 156.0 ppm; HRMS (ESI):
calcd for C13H16NaO4 m/z 259.0941 [M+Na]+, found m/z 259.0945
[M+Na]+.
2-(2-Phenethyl-1,3-dioxolan-2-yl)phenol (1a): 2’-Hydroxy-3-phe-
nylpropiophenone (1 g, 1 eq, 4.4 mmol), ethylene glycol (0.37 mL,
1.5 eq, 6.6 mmol), trimethylorthoformate (0.73 mL, 1.5 eq,
6.6 mmol), and trifluoromethanesulfonic acid (3.9 mL, 0.01 eq,
0.04 mmol) in hexane (7 mL) were heated at 428C o/n, then ethyl-
ene glycol (0.2 mL) and trimethylorthoformate (0.4 mL) were
added, and the reaction was heated for two more days. Then the
mixture was cooled to RT, and Et3N (3 mL) was added followed by
saturated aq NaHCO3 (25 mL). This was extracted with Et2O (4”
50 mL, a few drops Et3N added), and the combined organic ex-
tracts were washed once with brine (30 mL), dried over Na2SO4,
and concentrated to afford 1a as a white solid (99%, 1.17 g),
which was used without further purification. NMR data were in
good accordance with the literature.[25]
1-Amino-3-(2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)propan-2-
ol (3a): The epoxide 2a (2 g, 1 eq, 6.12 mmol), was dissolved in
MeOH (23 mL) and NH3 in MeOH (7m, 23 mL) was added. It was
stirred for 2 d at RT and then evaporated to yield a crude yellow oil
(2.28 g, 109 %), used without further purification; 1H NMR
(500 MHz, CDCl3): d=2.44 (t, J=8.5 Hz, 2H), 2.55–2.73 (m, 2H),
2.80–2.99 (m, 2H), 3.80–3.93 (m, 2H), 3.93–4.14 (m, 4H), 4.17–4.25
(m, 1H), 6.91 (d, J=6.8 Hz, 1H), 6.97 (t, J=6.8 Hz, 1H), 7.09–7.18
(m, 1H), 7.15 (d, J=7.0 Hz, 2H), 7.23 (d, J=6.3 Hz, 2H), 7.26–7.33
(m, 1H), 7.48 ppm (d, J=7.3 Hz, 1H); 13C NMR (125 MHz, CDCl3):
d=30.0, 40.0, 44.2, 64.7, 64.8, 70.8, 72.4, 110.3, 114.4, 120.0, 125.6,
127.1, 128.25 (2C), 128.30 (2C), 129.6, 130.0, 142.1, 156.6 ppm;
HRMS (ESI): calcd for C20H26NO4 m/z 344.1856 [M+H]+, found m/z
344.1857 [M+H]+.
5-Methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phenol (1b): Prepa-
ration like 1a; used without further purification. Brown oil (4.54 g,
97%); 1H NMR (500 MHz, CDCl3): d=2.20–2.28 (m, 2H), 2.70–2.79
(m, 2H), 3.79 (s, 3H), 3.94 (brs, 2H), 4.13 (brs, 2H), 6.41–6.51 (m,
2H), 7.13–7.23 (m, 4H), 7.23–7.36 (m, 2H), 8.37 ppm (s, 1H);
13C NMR (125 MHz, CDCl3): d=29.5, 41.4, 53.7, 64.5 (2C), 101.9,
106.3, 111.3, 117.5, 125.8, 127.8, 128.3 (2C), 128.3 (2C), 141.5, 155.9,
161.6 ppm.
1-Amino-3-(5-methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phe-
noxy)-propan-2-ol (3b): Prepared like 3a; yellow oil (1.70 g, 54%);
1H NMR (500 MHz, CDCl3): d=2.40 (t, J=8.7 Hz, 2H), 2.54–2.70(m,
2H), 2.83–2.97 (m, 2H), 3.79 (s, 3H), 3.82–3.90 (m, 2H), 3.92 (d, J=
6.3 Hz, 1H), 3.91–3.99 (m,1H), 4.00–4.10 (m, 2H), 4.17 (d, J=6.9 Hz,
1H), 6.44–6.52 (m, 2H), 7.09–7.18 (m, 3H), 7.20–7.26 (m, 2H),
2-(2-Methyl-1,3-dioxolan-2-yl)phenol (1c): Preparation like 1a. Pu-
rification by column chromatography (petrol ether:EtOAc, 20:1!
15:1) yielded 1c as a yellow oil (9.99 g, 63%). NMR data were in
good accordance with the literature.[16]
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1386 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
7.37 ppm (d, J=8.5 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=30.1,
40.2, 44.2, 55.4, 64.6, 64.7, 70.7, 72.4, 101.7, 105.1, 110.3, 122.5,
125.6, 127.9, 128.2 (2C), 128.3 (2C), 142.0, 157.5, 160.9 ppm; HRMS
(ESI): calcd for C21H28NO5 m/z 374.1962 [M+H]+, found m/z
374.1964 [M+H]+.
2-(5-Methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)acetoni-
trile (8b): Preparation like 8a. The crude product was purified by
column chromatography (petrol ether:EtOAc 4:1 to 2:1) to yield
8b as a yellow oil (3.83 g, 72%); 1H NMR (500 MHz, CDCl3): d=
2.34–2.41 (m, 2H), 2.59–2.66 (m, 2H), 3.82 (s, 3H), 3.85–3.91 (m,
2H), 4.05–4.09 (m, 2H), 4.79 (s, 2H), 6.61 (s, 1H), 6.62 (dd, J=6.5,
2.4 Hz, 1H), 7.12–7.17 (m, 3H), 7.21–7.26 (m, 2H), 7.48 ppm (dd, J=
6.3, 3.2 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=30.1, 40.2, 55.0,
55.5, 64.7 (2C), 102.9, 107.1, 125.6, 128.2 (2C), 128.3 (2C), 128.9,
109.5, 115.3, 124.0, 142.1, 155.1, 160.7 ppm; HRMS (ESI): calcd
for C20H21NNaO4 m/z 362.1363 [M+Na]+, found m/z 362.1366
[M+Na]+.
1-Amino-3-(2-(2-methyl-1,3-dioxolan-2-yl)phenoxy)propan-2-ol
(3c): Preparation like 3a; the crude product (mixture with dimer)
was purified by column chromatography (CH2Cl2 +1% Et3N to
CH2Cl2:MeOH=20:1+1% Et3N) to yield the pure product as a yel-
lowish solid (378 mg, 14%) and the rest (1.61 g) as a mixture with
1
the dimer; H NMR (500 MHz, CDCl3): d=1.77 (s, 3H), 2.82–2.92 (m,
2H), 3.77–3.88 (m, 2H), 3.92–3.99 (m, 2H), 4.01–4.08 (m, 2H), 4–17–
4.26 (m, 1H), 6.91 (d, J=8.2 Hz, 1H), 6.94 (td, J=7.6, 0.9 Hz, 1H),
7.23–7.29 (m, 1H), 7.46 ppm (dd, J=7.7, 1.7 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=25.8, 44.3, 64.4, 64.5, 71.1, 72.3, 108.8, 114.4,
120.9, 126.3, 129.5, 130.9, 156.5 ppm; HRMS (ESI): calcd for
C13H20NO4 m/z 254.1387 [M+H]+, found m/z 254.1390 [M+H]+.
2-(2-(2-Methyl-1,3-dioxolan-2-yl)phenoxy)acetonitrile (8c): Prepa-
ration like 8a. The crude product was purified by column chroma-
tography (petrol ether:EtOAc 4:1 to 1:1) to yield 8c as a white
solid (3.01 g, 81%);1H NMR (500 MHz, CDCl3): d=1.75 (s, 3H), 3.82–
3.86 (m, 2H), 4.04–4.08 (m, 2H), 4.85 (s, 2H), 7.08 (d, J=8.2 Hz,
1H), 7.10 (t, J=7.4 Hz, 1H), 7.34 (td, J=7.8, 1.7 Hz, 1H), 7.56 ppm
(dd, J=7.6, 1.6 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=25.8, 64.6
(2C), 55.1, 108.0, 114.5, 114.6, 123.4, 127.3, 129.7, 132.8, 154.3 ppm;
HRMS (ESI): calcd for C12H13NNaO3 m/z 242.0788 [M+Na]+, found
m/z 242.0789 [M+Na]+.
2-(2-(3-Phenylpropanoyl)phenoxy)acetonitrile (7a): 2’-Hydroxy-3-
phenylpropiophenone (4 g, 1 eq, 17.7 mmol) was dissolved in ace-
tone (200 mL), and K2CO3 (9.77 g, 4 eq, 70.7 mmol) was added. Bro-
moacetonitrile (2.54 g, 1.2 eq, 21.2 mmol) dissolved in acetone
(100 mL) was added dropwise throughout 30 min. The reaction
was stirred at 508C for 4 d, then it was filtered, washed once with
acetone, and concentrated. The residue was dissolved in CH2Cl2
(100 mL) and water (100 mL), and the aqueous phase was extract-
ed a CH2Cl2/CHCl3 mixture (10% CHCl3, 3”100 mL). The combined
organic extracts were washed once with brine (70 mL), dried over
Na2SO4, and the solvent was evaporated.to yield a brown solid
(4.15 g, 88%). NMR data were in good accordance with the litera-
ture.[18]
2-(2-(2-Phenethyl-1,3-dioxolan-2-yl)phenoxy)ethan-1-amine (9a):
Red-Al (~3.5m in toluene, 4.5 mL, 9.7 eq, 15.8 mmol) was placed in
a flask, and the nitrile 8a (0.5 g, 1 eq, 1.62 mmol) dissolved in tolu-
ene (20 mL) was added dropwise during 30 min. The reaction mix-
ture was stirred at RT for 3 h, and then water (20 mL) and Na-K-tar-
trate (13 g) was added and stirred for 30 min at RT. The mixture
was filtered over celite and washed with EtOAc. Then the phases
were separated, and the aqueous phase was extracted with EtOAc
(3”80 mL). The combined organic extracts were washed once with
brine (100 mL), dried over Na2SO4, and evaporated. The crude
product was purified by column chromatography (CH2Cl2 +1%
Et3N to CH2Cl2:MeOH=20:1+1% Et3N) to yield 9a as a colorless
oil (76 mg, 15%);1H NMR (500 MHz, CDCl3): d=2.43–2.50 (m, 2H),
2.61–2.69 (m, 2H), 3.08 (brs, 2H), 3.82–3.92 (m, 2H), 3.99–4.10 (m,
4H), 6.89 (d, J=8.2 Hz, 1H), 6.94 (t, J=7.4 Hz, 1H), 7.11–7.18 (m,
3H), 7.20–7.33 (m, 3H), 7.51 ppm (dd, J=7.6, 1.6 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=30.1, 39.7, 41.7, 64.7 (2C), 70.7, 110.0, 113.0,
120.2, 125.5, 127.4, 128.2 (2C), 128.3 (2C), 129.4, 129.7, 142.2,
156.5 ppm; HRMS (ESI): calcd for C19H24NO3 m/z 314.1751 [M+H]+,
found m/z 314.1752 [M+H]+.
2-(5-Methoxy-2-(3-phenylpropanoyl)phenoxy)acetonitrile (7b):
Preparation like 7a; brown solid (4.86 g, 93%);1H NMR (500 MHz,
CDCl3): d=3.02 (t, J=7.6 Hz, 2H), 3.24 (t, J=7.6 Hz, 2H), 3.87 (s,
3H), 4.80 (s, 2H), 6.52 (d, J=2.2 Hz, 1H), 6.66 (dd, J=8.7, 2.0 Hz,
1H), 7.17–7.22 (m, 1H), 7.22–7.25 (m, 2H), 7.27–7.32 (m, 2H),
7.81 ppm (d, J=8.8 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=30.4,
44.9, 54.0, 55.7, 100.4, 107.5, 114.5, 121.8, 126.0, 128.4 (4C), 133.1,
141.4, 157.0, 164.1, 198.5 ppm; HRMS (ESI): calcd for C18H17NNaO3
m/z 318.1101 [M+Na]+, found m/z 318.1100 [M+Na]+.
2-(2-Acetylphenoxy)acetonitrile (7c): Preparation like 7a. The
product was a white solid (3.54 g, 100%), which was used without
further purification. NMR data were in good accordance with the
literature.[26]
2-(5-Methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)ethan-1-
amine (9b): LiAlH4 (671 mg, 2 eq, 17.7 mmol) was suspended in
absolute Et2O (200 mL), and the nitrile 8b (3 g, 1 eq, 8.8 mmol, dis-
solved in 50 mL Et2O) was added carefully throughout 30 min.
After completion of the addition, the reaction mixture was heated
at reflux for 4 h and afterwards stirred at RT o/n. First EtOAc, and
then saturated Na-K-tartrate solution, was added carefully with
cooling, and the mixture was stirred for 3 h. Then the aqueous
phase was extracted with EtOAc (3”150 mL). The organic phases
were dried over Na2SO4, filtered, and evaporated. The crude prod-
uct was purified by column chromatography (CH2Cl2 to
CH2Cl2:MeOH 20:1, always 1% conc. aq NH4OH added) to yield
2-(2-(2-Phenethyl-1,3-dioxolan-2-yl)phenoxy)acetonitrile
(8a):
The nitrile 7a (4 g, 1 eq, 15.1 mmol), ethylene glycol (1.7 mL, 2 eq,
30.9 mmol), and p-toluenesulfonic acid monohydrate (0.29 g,
0.1 eq, 1.5 mmol) were dissolved/suspended in toluene (50 mL)
and heated o/n in a Dean–Stark apparatus. Then the reaction mix-
ture was poured into ice-cold sat. NaHCO3 solution (50 mL) and ex-
tracted with EtOAc (4”70 mL). The combined organic extracts
were dried over Na2SO4, filtered, and evaporated to yield a brown
1
1
oil (4.52 g, 97%); H NMR (500 MHz, CDCl3): d=2.35–2.44 (m, 2H),
a yellow oil (1.27 g, 42%); H NMR (500 MHz, CDCl3): d=2.40–2.46
2.60–2.69 (m, 2H), 3.84–3.93 (m, 2H), 4.04–4.13 (m, 2H), 4.81 (s,
2H), 7.07 (d, J=8.2 Hz, 1H), 7.09–7.18 (m, 4H), 7.20–7.26 (m, 2H),
7.35 (td, J=7.7, 1.7 Hz, 1H), 7.58 ppm (dd, J=7.7, 1.7 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=29.9, 40.0, 55.2, 64.9 (2C), 109.5,
115.5, 115.7, 123.3,125.6, 128.1, 128.2 (2C), 128.3 (2C), 129.8, 132.0,
142.0, 154.4 ppm; HRMS (ESI): calcd for C19H19NNaO3 m/z 332.1257
[M+Na]+, found m/z 332.1258 [M+Na]+.
(m, 2H), 2.59–2.66 (m, 2H), 3.07 (t, J=4.9 Hz, 2H), 3.80 (s, 3H),
3.83–3.88 (m, 2H), 4.01 (t, J=4.9 Hz, 2H), 4.02–4.06 (m, 2H), 6.43–
6.47 (m, 2H), 7.11–7.17 (m, 3H), 7.20–7.26 (m, 2H), 7.40 ppm (d, J=
8.8 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=30.2, 39.9, 41.6, 55.3,
64.6 (2C), 70.5, 100.7, 103.6, 110.0, 122.2, 125.5, 128.2 (3C), 128.3
(2C), 142.5, 157.4, 160.7 ppm; HRMS (ESI): calcd for C20H26NO4 m/z
344.1856 [M+H]+, found m/z 344.1860 [M+H]+.
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1387 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
2-(2-(2-Methyl-1,3-dioxolan-2-yl)phenoxy)ethan-1-amine
(9c):
N-(2-hydroxy-3-(5-methoxy-2-(3-phenylpropanoyl)phenoxy)-
propyl)-pyrrolidine-2-carboxamide (5b): Preparation of the pro-
tected derivative like 5a; purification by column chromatography
(petrol ether:EtOAc=1:1 to pure EtOAc to yield tert-butyl 2-((2-hy-
droxy-3-(5-methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)pro-
Preparation like 9b. The crude product was purified by column
chromatography (CH2Cl2 with 1% Et3N to CH2Cl2:MeOH=12:1+
1% Et3N) to yield a brown oil (744 mg, 39%); 1H NMR (500 MHz,
CDCl3): d=1.78 (s, 3H), 3.10 (t, J=5.0 Hz, 2H), 3.77–3.85 (m, 2H),
3.99–4.04 (m, 2H), 4.05 (t, J=5.0 Hz, 2H), 6.90 (d, J=8.2 Hz, 1H),
6.91 (td, J=7.5, 1.2 Hz, 1H), 7.25 (ddd, J=8.0, 7.4, 1.9 Hz, 1H),
7.49 ppm (dd, J=7.6, 1.6 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=
25.6, 41.7, 64.4 (2C), 70.7, 108.5, 113.0, 120.2, 126.7, 129.3, 130.6,
156.4 ppm; HRMS (ESI): calcd for C12H18NO3 m/z 224.1281 [M+H]+,
found m/z 224.1284 [M+H]+.
pyl)carbamoyl)pyrrolidine-1-carboxylate as
a
colorless honey
(220 mg, 71%); 1H NMR (500 MHz, CDCl3): d=1.80–1.91 (m, 2H),
1.32–1.47 (m, 9H), 2.04–2.24 (m, 2H), 2.38 (t, J=8.5 Hz, 2H)„ 2.54–
2.66 (m, 2H), 3.21–3.58 (m, 4H), 3.79 (s, 3H), 3.81–3.91 (m, 3H),
3.99–4.10 (m, 3H), 4.11–4.19 (m, 1H), 4.20–4.32 (m, 1H), 6.40–6.52
(m, 2H), 7.08–7.17 (m, 3H), 7.19–7.25 (m, 2H), 7.36 ppm (d, J=
8.5 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=24.6, 28.2/28.3 (3C),
30.1/30.2, 31.2, 40.2, 41.4/41.5, 47.1, 55.4, 61.3, 64.6, 64.7, 68.9, 71.9,
80.1/80.7, 101.7/101.8, 105.3, 110.6, 122.4, 125.6, 127.9, 128.2 (2C),
128.3 (2C), 142.1, 157.4, 160.9, 173.3 ppm; HRMS (ESI): calcd for
C31H42N2NaO8 m/z 593.2833 [M+Na]+, found m/z 593.2838
[M+Na]+.
N-(2-hydroxy-3-(2-(3-phenylpropanoyl)phenoxy)propyl)pyrroli-
dine-2-carboxamide (5a): Coupling reaction: The amine 4a
(1000 mg, 1 eq, 2.91 mmol) and boc-proline (627 mg, 1 eq,
2.91 mmol) were dissolved in anhydrous CH2Cl2 (15 mL), and dime-
thylaminopyridine (18 mg, 0.05 eq, 0.15 mmol) and dicyclohexylcar-
bodiimide (601 mg, 1 eq, 2.91 mmol) were added and stirred until
TLC indicated completion (2 h to o/n). The reaction mixture was fil-
tered, washed with sat. NaHCO3 solution (20 mL) and water
(20 mL), and the aqueous extracts were extracted with CH2Cl2 (2”
50 mL). The combined organic extracts were washed once with
brine (50 mL), dried over Na2SO4, and the solvent was evaporated.
The product was purified by column chromatography (CH2Cl2 to
CH2Cl2:MeOH=30:1 to 15:1) to yield tert-butyl 2-((2-hydroxy-3-(2-
(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)propyl)carbamoyl)pyrroli-
dine-1-carboxylate as a white solid (mixture of diastereomers,
Deprotection like for 5a; orange honey (134 mg, 99%); 1H NMR
(500 MHz, CDCl3): d=1.73 (quin, J=6.9 Hz, 2H), 1.86–1.96 (m, 1H),
2.12–2.22 (m, 1H), 2.87–2.98 (m, 1H), 2.98–3.06 (m, 1H), 3.02 (t, J=
7.6 Hz, 2H), 3.25 (t, J=7.7 Hz, 2H), 3.30–3.39 (m, 1H), 3.49–3.61 (m,
1H), 3.81–3.87 (m, 1H), 3.84 (s, 3H), 3.97–4.04 (m, 2H), 4.04–4.11
(m, 1H), 6.46 (d, J=1.9 Hz, 1H), 6.53 (d, J=8.8 Hz, 1H), 7.19 (t, J=
7.3 Hz, 1H), 7.23 (d, J=7.3 Hz, 2H), 7.26–7.34 (m, 2H), 7.74 (d, J=
8.8 Hz, 1H), 8.05 ppm (vrs, 1H); 13C NMR (125 MHz, CDCl3): d=
26.1/26.2, 30.4, 30.8, 42.9, 44.5, 47.2/47.3, 55.6, 60.3, 69.8, 70.6/70.7,
99.9/100.0, 105.9/106.0, 121.0, 126.0, 128.3/128.4 (2C), 128.4/128.5
(2C), 132.5, 141.6, 159.8, 164.3, 177.4, 199.2 ppm; HRMS (ESI): calcd
for C24H30N2NaO5 m/z 449.2047 [M+Na]+, found m/z 449.2050
[M+Na]+.
1
986 mg, 63%); H NMR (500 MHz, CDCl3): d=1.41 (s, 9H), 1.86 (brs,
2H), 2.07–2.31 (m, 2H), 2.42 (t, J=8.4 Hz, 2H), 2.55–2.70 (m, 2H),
3.34 (brs, 1H), 3.39–3.51 (m, 2H), 3.55 (brs, 1H), 3.80–3.94 (m, 2H),
4.00–4.12 (m, 2H), 4.16–4.31 (m, 2H), 6.90 (d, J=7.9 Hz, 1H), 6.96
(t, J=7.9 Hz, 1H), 7.11–7.17 (m, 3H), 7.20–7.29 (m, 3H), 7.47 ppm
(d, J=7.9 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=22.1/23.7,
28.2(3C), 30.0, 31.1, 40.0, 41.5, 47.0/47.1, 61.3, 64.7, 64.8, 69.0/69.3,
71.6/71.9, 79.9, 110.4, 114.6, 121.0/121.2, 125.6, 127.0, 128.3 (4C),
129.7, 130.0, 142.0, 156.5, 172.1 ppm; HRMS (ESI): calcd for
C30H40N2NaO7 m/z 563.2728 [M+Na]+, found m/z 563.2733
[M+Na]+.
Hydrochloride formation like for 5a yielded a brown honey
(108 mg, 98%); HRMS (ESI): calcd for C24H30N2NaO5 m/z 449.2047
[M+Na]+, found m/z 449.2047 [M+Na]+.
N-(3-(2-acetylphenoxy)-2-hydroxypropyl)pyrrolidine-2-carboxa-
mide (5c): Preparation of the protected derivative like for 5a. The
product was purified by column chromatography (CH2Cl2 to
CH2Cl2:MeOH=30:1) to yield tert-butyl2-((2-hydroxy-3-(2-(2-methyl-
1,3-dioxolan-2-yl)phenoxy)propyl)carbamoyl)pyrrolidine-1-carboxyl-
ate as a colorless semisolid (109 mg, 41%); 1H NMR (500 MHz,
CDCl3): d=1.38 (brs, 6H), 1.42 (brs, 3H), 1.76 (s, 3H), 1.79–1.90 (m,
2H), 1.95–2.27 (m, 2H), 3.22–3.41 (m, 1H), 3.41–3.51 (m, 2H), 3.52–
3.73 (m, 1H), 3.76–3.95 (m, 3H), 4.00–4.12 (m, 3H), 4.15–4.32 (m,
2H), 6.84 (brs, 0.5H), 6.90 (d, J=7.9 Hz, 1H), 6.94 (t, J=7.7 Hz, 1H),
7.14 (brs, 0.5H), 7.26 (t, J=7.7 Hz, 1H), 7.45 ppm (d, J=7.6 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=23.7/24.5, 25.8, 28.2/28.3 (3C), 31.1,
41.5/41.6, 47.0/47.2, 60.3/61.2, 64.4/64.5, 69.0/69.4, 71.5/71.8, 108.9,
114.5, 121.1, 126.2, 126.3, 129.6, 130.9, 156.4, 175.7 ppm; HRMS
(ESI): calcd for C23H34N2NaO7 m/z 473.2258 [M+Na]+, found m/z
473.2255 [M+Na]+.
Deprotection: The protected derivative (500 mg, 1 eq, 0.92 mmol)
was dissolved in dioxane (3 mL) and aq. HCl (4m, 5 mL) was added
and stirred at RT for 4 h. Then the reaction mixture was basified
with aq NaOH solution, diluted with water, and extracted with
EtOAc (4”50 mL), dried over Na2SO4; and the solvent was evapo-
rated to yield a brown oil (380 mg, 98%); 1H NMR (500 MHz,
CDCl3): d=1.63–1.73 (m, 2H), 1.83–1.93 (m, 2H), 2.04–2.19 (m, 1H),
2.74–2.94 (m 2H), 3.03 (t, J=7.6 Hz, 2H), 3.28 (t, J=7.7 Hz, 2H),
3.30–3.39 (m, 1H), 3.46–3.55 (m, 1H), 3.70–3.78 (m, 1H), 3.94–4.08
(m, 3H), 6.95 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H), 7.18 (t, J=
7.3 Hz, 1H), 7.22 (d, J=7.0 Hz, 2H), 7.28 (d, J=8.2 Hz, 2H), 7.43 (t,
J=7.9 Hz, 1H), 7.60 (d, J=7.6 Hz, 1H), 8.02 ppm (brs, 1H); 13C NMR
(125 MHz, CDCl3): d=26.1, 30.1, 30.7/30.8, 42.7/42.8, 44.7, 47.2,
60.3, 69.8, 70.5/70.6, 113.3/113.4, 121.1, 126.0, 128.3 (2C), 128.4
(2C), 129.9/130.0, 133.4, 141.3, 157.3, 177.3, 201.6 ppm; HRMS (ESI):
calcd for C23H29N2NaO4 m/z 419.1941 [M+Na]+, found m/z
419.1939 [M+Na]+.
Deprotection like for 5a yielded a yellowish semi-solid (50 mg,
96%); 1H NMR (500 MHz, CDCl3): d=1.62–1.78 (m,2H), 1.81–1.94
(m, 1H), 2.08–2.21 (m, 1H), 2.61 (s, 3H), 2.81–2.95 (m, 1H), 2.95–
3.06 (m, 1H), 3.38–3.49 (m, 1H), 3.56–3.66 (m, 1H), 3.72–3.87 (m,
1H), 3.97–4.09 (m, 2H), 4.09–4.19 (m, 1H), 6.96 (d, J=8.5 Hz, 1H),
7.01 (t, J=7.6 Hz, 1H), 7.44 (t, J=7.7 Hz, 1H), 7.68 (d, J=7.6 Hz,
1H), 8.13 ppm (brs, 1H); 13C NMR (125 MHz, CDCl3): d=26.1, 30.8,
31.2, 42.8, 47.2, 60.3, 69.8, 70.6/70.7, 113.3, 121.0, 128.3, 130.4,
133.7, 157.7, 177.3, 199.9 ppm; HRMS (ESI): calcd for C16H23N2O4 m/
z 307.1652 [M+H]+, found m/z 307.1657 [M+H]+.
Hydrochloride formation: The amine (320 mg, 0.8 mmol) was dis-
solved in 3 mL methyl tert-butyl ether (MTBE) and CH3CN (1 mL)
and treated with 6n HCl (1 mL), stirred for 1 h, and then the sol-
vent was evaporated to yield the hydrochloride as a yellow semi-
solid (347 mg, 87%)’; HRMS (ESI): calcd for C23H29N2NaO4 m/z
419.1941 [M+Na]+, found m/z 419.1939 [M+Na]+.
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1388 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Hydrochloride formation like for 5a yielded a brown honey
(60 mg, 94%); HRMS (ESI): calcd for C16H23N2O4 m/z 307.1652
[M+H]+, found m/z 307.1657 [M+H]+.
Hydrochloride formation like for 5a yielded a brown honey
(286 mg, 99%); HRMS (ESI): calcd for C23H29N2O4 m/z 397.2122
[M+H]+, found m/z 397.2124 [M+H]+.
N-(2-(2-acetylphenoxy)ethyl)pyrrolidine-2-carboxamide
(10c):
N-(2-(2-(3-phenylpropanoyl)phenoxy)ethyl)pyrrolidine-2-carbox-
amide (10a): Preparation like 5a; purification by column chroma-
tography (petrol ether:EtOAc=2:1 to 1:2) led to tert-butyl 2-((2-(2-
(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)ethyl)carbamoyl)pyrroli-
dine-1-carboxylate as a yellowish honey (85 mg, 62%); 1H NMR
(500 MHz, CDCl3): d=1.33 (brs, 6H), 1.42 (brs, 3H), 1.73–1.92 (m,
2H), 1.95–2.02 (m, 1H), 2.11–2.22 (m, 1H), 2.38–2.52 (m, 2H), 2.55–
2.67 (m, 2H), 3.42–3.55 (m, 2H), 3.61–3.72 (m, 2H), 3.81–3.95 (m,
2H), 4.02–4.26 (m, 5H), 6.90 (d, J=8.2 Hz, 1H), 6.97 (t, J=7.4 Hz,
1H), 7.08–7.17 (m, 3H), 7.19–7.25 (m, 2H), 7.28 (t, J=7.6 Hz, 1H),
7.50 ppm (d, J=7.6 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=23.8,
28.3 (3C), 30.2, 31.6, 39.0, 39.7, 46.8, 61.2, 64.7, 64.8, 67.9, 80.0,
110.3, 113.7, 120.9, 125.6, 127.6, 128.2 (4C), 129.6, 129.8, 142.1,
156.0, 173.3 ppm; HRMS (ESI): calcd for C29H39N2O6 m/z 511.2803
[M+H]+, found m/z 511.2807 [M+H]+.
Preparation like 5a, purification by column chromatography
(CH2Cl2:MeOH=40:1 to 25:1) yielded tert-butyl 2-((2-(2-(2-methyl-
1,3-dioxolan-2-yl)phenoxy)ethyl)carbamoyl)pyrrolidine-1-carboxyl-
1
ate as a colorless oil (215 mg, 57%); H NMR (500 MHz, CDCl3): d=
1.36 (brs, 6H), 1.43 (brs, 3H), 1.77 (s, 3H), 1.81–1.96 (m, 2H), 1.97–
2.07 (m, 1H), 2.16–2.25 (m, 1H), 3.44–3.57 (m, 2H), 3.64–3.74 (m,
2H), 3.77–3.89 (m, 2H), 4.03–4.18 (m, 4.5H), 4.22–4.29 (m, 0.5H),
6.91 (d, J=8.2 Hz, 1H), 6.95 (t, J=7.1 Hz, 1H), 7.27 (t, J=7.1 Hz,
1H), 7.48 ppm (d, J=7.9 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=
23.8, 25.6, 28.3(3C), 31.5, 39.0, 46.9, 61.3, 64.4 (2C), 68.1, 80.2, 108.7,
114.1, 121.1, 126.8, 129.5, 131.1, 155.9, 165.7 ppm; HRMS (ESI):
calcd for C22H33N2O6 m/z 421.2333 [M+H]+, found m/z 421.2338
[M+H]+.
Deprotection like 5a yielded a yellowish solid (75 mg, 97%);
1H NMR (500 MHz, CDCl3): d=1.66–1.77 (m, 2H), 1.85–1.97 (m, 1H),
2.11–2.23 (m, 1H), 2.64 (s, 3H), 2.84–2.94 (m, 1H), 2.98–3.07 (m,
1H), 3.62–3.77 (m, 2H), 3.78–3.85 (m, 1H), 4.16 (t, J=4.7 Hz, 2H),
6.94 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H), 7.45 (t, J=7.9 Hz,
1H), 7.73 (d, J=7.9 Hz, 1H), 8.13 ppm (brs, 1H); 13C NMR (125 MHz,
CDCl3): d=26.1, 30.6, 31.8, 38.5, 47.1, 60.4, 67.3, 112.6, 121.0, 128.5,
130.5, 133.7, 157.8, 170.7, 199.7 ppm; HRMS (ESI): calcd for
C15H21N2NaO3 m/z 299.1366 [M+Na]+, found m/z 299.1368
[M+Na]+.
Deprotection like for 5a; yellow honey (48 mg, quant.); 1H NMR
(500 MHz, CDCl3): d=1.62 (quin, J=6.9 Hz, 2H), 1.78–1.86 (m, 1H),
2.05–2.14 (m, 1H), 2.69–2.78 (m, 1H), 2.84–2.91 (m, 1H), 3.05 (t, J=
7.3 Hz, 2H), 3.27–3.40 (m, 2H), 3.56–3.74 (m, 3H), 4.08–4.19 (m,
2H), 6.93 (d, J=8.5 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H), 7.19 (t, J=
7.3 Hz, 1H), 7.23 (d, J=7.0 Hz, 1H), 7.26–7.32 (m, 2H), 7.44 (ddd,
J=8.7, 7.1, 1.6 Hz, 1H), 7.66 (dd, J=7.9, 1.6 Hz, 1H), 7.98 ppm (brs,
1H); 13C NMR (125 MHz, CDCl3): d=26.2, 30.3, 30.6, 38.4, 45.3, 47.2,
60.4, 67.3, 112.4, 121.1, 126.0, 128.4 (4C), 130.3, 132.3, 133.4, 141.7,
157.5, 174.2, 201.8 ppm; HRMS (ESI): calcd for C22H27N2O3 m/z
367.2016 [M+H]+, found m/z 367.2019 [M+H]+.
Hydrochloride formation like for 5a yielded a brown semisolid
(86 mg, 98%); HRMS (ESI): calcd for C15H21N2NaO3 m/z 299.1366
[M+Na]+, found m/z 299.1368 [M+Na]+.
2-(2-Hydroxy-3-(2-(3-phenylpropanoyl)phenoxy)propyl)-hexahy-
dropyrrolo[1,2-a]pyrazine-1,4-dione (6a): The amine 4a (371 mg,
1 eq, 1.08 mmol), methyl (2-chloroacetyl)prolinate (223 mg, 1 eq,
1.08 mmol), and Et3N (0.16 mL, 1.2 eq, 1.13 mmol) were dissolved
in CH3CN (40 mL) and stirred at RT for 2.5 d. The solvent was
evaporated, and the residue was redissolved in EtOAc and extract-
ed with water. The aqueous phase was extracted with EtOAc (3”
50 mL), the combined organic phases were dried over Na2SO4, and
the solvent was evaporated. The product was purified by column
chromatography (CH2Cl2 to CH2Cl2:MeOH=20:1) twice to yield
a yellow semisolid (30 mg, 6%).
Hydrochloride formation like for 5a yielded a brown semisolid
(47 mg, 98%); HRMS (ESI): calcd for C22H27N2O3 m/z 367.2016
[M+H]+, found m/z 37.2020 [M+H]+.
N-(2-(4-methoxy-2-(3-phenylpropanoyl)phenoxy)ethyl)pyrroli-
dine-2-carboxamide (10b): Preparation like 5a; purification by
column chromatography (petrol ether:EtOAc=1:3 to 1:5) led to
tert-butyl
2-((2-(5-methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phe-
noxy)ethyl)carbamoyl)pyrrolidine-1-carboxylate as a colorless oil
(989 mg, 78%); 1H NMR (500 MHz, CDCl3): d=1.33 (brs, 6H), 1.42
(brs, 3H), 1.75–1.84 (m, 1H), 1.84–1.92 (m, 1H), 1.92–2.03 (m, 1H),
2.11–2.22 (m, 1H), 2.34–2.47 (m, 2H), 2.52–2.66 (m, 2H), 3.41–3.55
(m, 2H), 3.62–3.70 (m, 2H), 3.80 (s, 3H), 3.82–3.94 (m, 2H), 3.99–
4.23 (m, 5H), 6.46 (s, 1H), 6.48 (d, J=8.8 Hz, 1H), 7.09–7.17 (m,
3H), 7.20–7.25 (m, 2H), 7.40 ppm (d, J=8.2 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=23.8, 28.3 (3C), 30.3, 31.6, 38.8, 39.9, 46.8,
55.4, 61.2, 64.6, 64.7, 67.9, 80.1, 101.2, 104.6, 110.2, 122.3, 125.6,
128.2 (4C), 128.4, 142.3, 156.7, 160.8, 171.3 ppm; HRMS (ESI): calcd
for C30H41N2O7 m/z 541.2908 [M+H]+, found m/z 541.2911 [M+H]+.
2-(2-Hydroxy-3-(2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)propyl)-
1
hexahydropyrrolo[1,2-a]pyrazine-1,4-dione: Diastereomer 1: H NMR
(500 MHz, CDCl3): d=1.85–1.96 (m, 1H), 2.00–2.14 (m, 2H), 2.35–
2.43 (m, 1H), 2.41 (t, J=8.7 Hz, 2H), 2.60–2.68 (m, 2H), 3.26 (dd,
J=14.1, 7.4 Hz, 1H), 3.51–3.59 (m, 1H), 3.59–3.67 (m, 1H), 3.82–
3.92 (m, 4H), 4.05–4.13 (m, 3H), 4.04 (d, J=17.0 Hz, 1H), 4.16–4.19
(m, 1H), 4.23 (dd, J=9.5, 3.2 Hz, 1H), 4.42 (d, J=16.7 Hz, 1H), 6.90
(dd, J=8.2, 2.8 Hz, 1H), 6.98 (t, J=7.6 Hz, 1H), 7.10–7.17 (m, 3H),
7.20–7.26 (m, 2H), 7.28 (t, J=7.7 Hz, 1H), 7.48 ppm (d, J=7.6 Hz,
1H); 13C NMR (125 MHz, CDCl3): d=22.7, 28.9, 30.0, 40.1, 45.2, 49.7,
54.5, 59.1, 64.8 (2C), 69.7, 72.1, 110.4, 114.7, 121.3, 125.6, 127.1,
128.3 (4C), 129.8, 130.0, 142.0, 156.5, 163.6, 168.0 ppm; Diastereo-
Deprotection like for 5a yielded a yellow semisolid (356 mg, 98%);
1H NMR (500 MHz, CDCl3): d=1.56–1.64 (m, 2H), 1.76–1.86 (m, 1H),
2.01–2.12 (m, 1H), 2.67–2.75 (m, 1H), 2.81–2.89 (m, 1H), 3.03 (t, J=
7.7 Hz, 2H), 3.21–3.38 (m, 2H), 3.53–3.59 (m, 1H), 3.59–3.68 (m,
2H), 3.84 (s, 3H), 4.05–4.18 (m, 2H), 6.42 (s, 1H), 6.54 (d, J=8.8 Hz,
1H), 7.18 (t, J=7.1 Hz, 1H), 7.23 (d, J=7.3 Hz, 2H), 7.28 (d, J=
7.6 Hz, 2H), 7.81 (d, J=8.5 Hz, 1H), 7.99 ppm (brs, 1H); 13C NMR
(125 MHz, CDCl3): d=26.1, 30.4, 30.6, 38.3, 45.2, 47.2, 55.5, 60.4,
67.3, 98.9, 105.7, 121.0, 125.9, 128.3 (2C), 128.4 (2C), 132.7, 141.9,
159.6, 164.3, 175.6, 199.1 ppm; HRMS (ESI): calcd for C23H29N2O4 m/
z 397.2122 [M+H]+, found m/z 397.2126 [M+H]+.
1
mer 2: H NMR (500 MHz, CDCl3): d=1.85–1.96 (m, 1H), 2.00–2.14
(m, 2H), 2.35–2.43 (m, 1H), 2.41 (t, J=8.7 Hz, 2H), 2.60–2.68 (m,
2H), 3.51–3.59 (m, 1H), 3.59–3.67 (m, 3H), 3.82–3.92 (m, 3H), 4.05–
4.13 (m, 3H), 4.14 (d, J=17.0 Hz, 1H), 4.18–4.22 (m, 2H), 4.29 (d,
J=16.4 Hz, 1H), 6.90 (dd, J=8.2, 2.8 Hz, 1H), 6.98 (t, J=7.6 Hz,
1H), 7.10–7.17 (m, 3H), 7.20–7.26 (m, 2H), 7.28 (t, J=7.7 Hz, 1H),
7.48 ppm (d, J=7.6 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=22.6,
28.9, 30.0, 40.0, 45.2, 48.9, 54.0, 58.9, 64.8 (2C), 69.5, 72.0, 110.3,
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1389 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
114.7, 121.2, 125.6, 127.1, 128.3 (4C), 129.7, 130.0, 142.0, 156.5,
163.4, 168.0 ppm; HRMS (ESI): calcd for C27H32N2NaO6 m/z 503.2153
[M+Na]+, found m/z 503.2151 [M+Na]+.
55.6, 59.0, 69.4, 71.6, 101.0, 106.1, 126.0, 128.4 (2C), 128.5 (2C),
132.5, 141.5, 159.8, 163.3, 164.2, 168.7, 199.2 ppm; Diastereomer 2:
1H NMR (500 MHz, CDCl3): d=1.86–1.97 (m, 1H), 2.00–2.12 (m, 2H),
2.34–2.43 (m, 1H), 3.02 (t, J=7.6 Hz, 1H), 3.19–3.25 (m, 2H), 3.52
(dd, J=14.2, 7.6 Hz, 1H), 3.54–3.65 (m, 2H), 3.68 (dd, J=14.3,
3.0 Hz, 1H), 3.84 (s, 3H), 3.91–4.03 (m, 1H), 3.98 (d, J=17.1 Hz, 1H),
4.07–4.15 (m, 2H), 4.15–4.21 (m, 1H), 4.21 (d, J=17.0 Hz, 1H), 6.45
(d, J=2.2 Hz, 1H), 6.53 (dd, J=8.7, 2.0 Hz, 1H), 7.18 (dd, J=8.1,
5.9 Hz, 1H), 7.23 (d, J=7.3 Hz, 2H), 7.28 (td, J=7.6, 1.7 Hz, 2H),
7.57 ppm (t, J=8.4 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=22.6,
28.8, 30.4, 44.0, 45.2, 49.7, 53.9, 55.6, 58.9, 69.2, 71.5, 100.9, 106.0,
126.0, 128.4 (2C), 128.5 (2C), 132.4, 141.5, 159.8, 163.2, 164.2, 168.5,
199.1 ppm; HRMS (ESI): calcd for C26H30N2NaO6 m/z 489.1996
[M+Na]+, found m/z 489.2004 [M+Na]+.
The ketal (24 mg, 1 eq, 0.05 mmol) was dissolved in CH3CN (2 mL),
and aq. HCl solution (2 N, 2 mL) was added, and it was stirred for
4 h at RT. Then it was diluted with water and extracted with EtOAc
(4”50 mL), dried over Na2SO4, and evaporated to yield a colorless
1
solid (21 mg, 96%). Diastereomer 1: H NMR (500 MHz, CDCl3): d=
1.86–1.97 (m, 1H), 2.00–2.12 (m, 2H), 2.35–2.43 (m, 1H), 3.04 (t, J=
7.9 Hz, 2H), 3.23–3.29 (m, 2H), 3.31 (dd, J=14.5, 7.9 Hz, 1H), 3.54–
3.65 (m, 2H), 3.79 (dd, J=14.1, 3.0 Hz, 1H), 3.93–4.01 (m, 1H), 3.96
(d, J=16.7 Hz, 1H), 4.07–4.15 (m, 2H), 4.17–4.22 (m, 1H), 4.33 (d,
J=17.0 Hz, 1H), 6.96 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H),
7.16–7.22 (m, 1H), 7.23 (d, J=7.3 Hz, 2H), 7.28 (td, J=7.5, 1.6 Hz,
2H), 7.44 (t, J=7.9 Hz, 1H), 7.60 ppm (ddd, J=7.7, 4.6, 1.7 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=22.6, 28.8, 30.1, 44.4, 45.2, 50.2, 54.4,
59.0, 69.5, 71.5, 114.3, 121.4, 126.1, 128.4 (2C), 128.5, 128.5 (2C),
130.0, 133.5, 141.3, 157.3, 163.4, 168.7, 201.5 ppm; Diastereomer 2:
1H NMR (500 MHz, CDCl3): d=1.86–1.97 (m, 1H), 2.00–2.12 (m, 2H),
2.35–2.43 (m, 1H), 3.04 (t, J=7.9 Hz, 2H), 3.23–3.29 (m, 2H), 3.52
(dd, J=14.4, 3.3 Hz, 1H), 3.54–3.65 (m, 2H), 3.68 (dd, J=14.2,
6.6 Hz, 1H), 3.93–4.01 (m, 1H), 4.01 (d, J=16.7 Hz, 1H), 4.07–4.15
(m, 2H), 4.16–4.20 (m, 1H), 4.23 (d, J=16.7 Hz, 1H), 6.96 (d, J=
8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H), 7.16–7.22 (m, 1H), 7.23 (d, J=
7.3 Hz, 2H), 7.28 (td, J=7.5, 1.6 Hz, 2H), 7.44 (t, J=7.9 Hz, 1H),
7.60 ppm (ddd, J=7.7, 4.6, 1.7 Hz, 1H); 13C NMR (125 MHz, CDCl3):
d=22.6, 28.8, 30.1, 44.3, 45.2, 49.6, 53.9, 58.9, 69.3, 71.4, 114.2,
121.4, 126.1, 127.0, 128.4 (2C), 128.5, 128.5 (2C), 133.5, 141.2, 157.3,
163.3, 168.5, 201.4 ppm; HRMS (ESI): calcd for C25H28N2NaO5 m/z
459.1890 [M+Na]+, found m/z 459.1893 [M+Na]+.
2-(3-(2-Acetylphenoxy)-2-hydroxypropyl)hexahydropyrrolo[1,2-
a]-pyrazine-1,4-dione (6c): Preparation like 6b. The product was
purified by column chromatography (CH2Cl2 to CH2Cl2:MeOH=
30:1) to yield a colorless oil (81 mg, yield given over two steps
below) that was already partially deprotected: 2-(2-hydroxy-3-(2-(2-
methyl-1,3-dioxolan-2-yl)phenoxy)propyl)
hexahydropyrrolo[1,2-
a]pyrazine-1,4-dione; HRMS (ESI): calcd for C20H26N2NaO6 m/z
413.1683 [M+Na]+, found m/z 413.1679 [M+Na]+.
Deprotection like 6b yielded a yellowish solid (23 mg, 35%, 2
1
steps). Diastereomer 1: H NMR (500 MHz, CDCl3): d=1.87–1.97 (m,
1H), 2.00–2.14 (m, 2H), 2.36–2.44 (m, 1H), 2.62 (s, 3H), 3.57 (dd, J=
13.9, 3.5 Hz, 1H), 3.53–3.59 (m, 1H), 3.59–3.66 (m, 1H), 3.80 (dd, J=
14.3, 6.8 Hz, 1H), 3.97–4.04 (m, 1H), 4.07 (d, J=16.7 Hz, 1H), 4.10–
4.19 (m, 2H), 4.22–4.31 (m, 1H), 4.34 (d, J=16.1 Hz, 1H), 6.96 (d,
J=8.2 Hz, 1H), 7.05 (t, J=7.4 Hz, 1H), 7.46 (t, J=7.9 Hz, 1H),
7.70 ppm (ddd, J=7.7, 2.0, 1.7 Hz, 1H); 13C NMR (125 MHz, CDCl3):
d=22.7, 28.8, 30.8, 45.2, 49.6, 54.0, 59.0, 69.5, 71.4, 114.2, 121.4,
130.6, 133.8, 128.5, 157.4, 163.2, 168.6, 198.2 ppm; Diastereomer 2:
1H NMR (500 MHz, CDCl3): d=1.87–1.97 (m, 1H), 2.00–2.14 (m, 2H),
2.36–2.44 (m, 1H), 2.61 (s, 3H), 3.42 (dd, J=14.2, 7.9 Hz, 1H), 3.53–
3.59 (m, 1H), 3.59–3.66 (m, 1H), 3.85 (dd, J=14.2, 3.2 Hz, 1H),
3.97–4.04 (m, 1H), 4.03 (d, J=17.0 Hz, 1H), 4.10–4.19 (m, 2H),
4.22–4.31 (m, 1H), 4.42 (d, J=16.4 Hz, 1H), 6.96 (d, J=8.2 Hz, 1H),
7.05 (t, J=7.4 Hz, 1H), 7.46 (t, J=7.9 Hz, 1H), 7.70 ppm (ddd, J=
7.7, 2.0, 1.7 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=22.7, 28.8, 30.7,
45.2, 50.4, 54.5, 58.9, 69.3, 71.3, 114.1, 121.4, 130.6, 133.8, 128.5,
157.4, 163.2, 168.6, 198.2 ppm; HRMS (ESI): calcd for C18H22N2NaO5
m/z 369.1421 [M+Na]+, found m/z 369.1427 [M+Na]+.
2-(2-Hydroxy-3-(5-methoxy-2-(3-phenylpropanoyl)phenoxy)prop-
yl)-hexahydropyrrolo[1,2-a]pyrazine-1,4-dione (6b): The amine
4b (300 mg, 1 eq, 0.8 mmol), methyl (2-chloroacetyl)prolinate
(165 mg, 1 eq, 0.8 mmol), and Et3N (0.14 mL, 1.3 eq, 1.0 mmol)
were dissolved in 2-ethoxyethanol (6 mL) and heated at reflux for
3 d. The solvent was evaporated and the residue was redissolved
in a mixture of CH2Cl2 and toluene and extracted with water. The
aqueous phase was extracted with CH2Cl2 (3”60 mL), the com-
bined organic phases were washed once with brine (70 mL), dried
over Na2SO4 and the solvent was evaporated. The product was pu-
rified by column chromatography (CH2Cl2 to CH2Cl2:MeOH=30:1)
to yield 251 mg of a yellow solid that was already partially depro-
tected.
2-(2-(2-(3-Phenylpropanoyl)phenoxy)ethyl)hexahydropyrro-
lo[1,2-a]-pyrazine-1,4-dione (11a): Preparation like 6b yielded
a mixture of partially deprotected 2-(2-(2-(2-phenethyl-1,3-dioxo-
lan-2-yl)phenoxy)ethyl)hexahydropyrrolo[1,2-a]pyrazine-1,4-dione
(114 mg, partially deprotected, therefore yield given for two steps
below); HRMS (ESI): calcd for C26H30N2NaO5 m/z 473.2047 [M+Na]+,
found m/z 473.2050 [M+Na]+.
2-(2-hydroxy-3-(5-methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phe-
noxy)propyl)hexahydropyrrolo[1,2-a]pyrazine-1,4-dione: HRMS (ESI):
calcd for C28H34N2NaO7 m/z 533.2258 [M+Na]+, found m/z
533.2252 [M+Na]+.
The mixture from the previous step (120 mg, 1 eq, 0.07 mmol) was
dissolved in CH3CN (9 mL); aq. HCl solution (2 N, 9 mL) was added
and it was stirred for 4 h at RT. Then it was diluted with water
(5 mL) and aq. NaHCO3, (10 mL), extracted with EtOAc (4”30 mL),
dried over Na2SO4, and evaporated to yield a yellowish solid
Deprotection like 6b yielded a yellowish semisolid (80 mg, 39%, 2
1
steps); H NMR (500 MHz, CDCl3): d=1.82–1.94 (m, 1H), 1.97–2.08
1
(109 mg, 61%, 2 steps). Diastereomer 1: H NMR (500 MHz, CDCl3):
(m, 2H), 2.29–2.37 (m, 1H), 3.02 (t, J=7.7 Hz, 2H), 3.23–3.29 (m,
2H), 3.49–3.61 (m, 3H), 3.86 (d, J=16.4 Hz, 1H), 3.94 (t, J=7.7 Hz,
1H), 4.02 (ddd, J=14.3, 5.8, 4.3 Hz, 1H), 4.15 (ddd, J=10.1, 5.7,
4.4 Hz, 1H), 4.23 (d, J=16.4 Hz, 1H) 4.23–4.29 (m, 1H), 6.92 (d, J=
8.2 Hz, 1H), 7.01 (t, J=7.6 Hz, 1H), 7.18 (t, J=7.1 Hz, 1H), 7.23 (d,
J=7.0 Hz, 2H), 7.26–7.30 (m, 2H), 7.42 (ddd, J=8.4, 7.4, 1.7 Hz,
1H), 7.57 ppm (dd, J=7.7, 1.7 Hz, 1H); 13C NMR (125 MHz, CDCl3):
d=22.6, 28.6, 30.2, 45.0, 45.2, 46.1, 53.4, 58.9, 65.8, 112.5, 121.3,
126.0, 128.4 (4C), 129.0, 130.0, 133.1, 141.4, 156.6, 163.0, 167.8,
d=1.86–1.97 (m, 1H), 2.00–2.12 (m, 2H), 2.34–2.43 (m, 1H), 3.02 (t,
J=7.6 Hz, 1H), 3.19–3.25 (m, 2H), 3.30 (dd, J=14.2, 7.6 Hz, 1H),
3.54–3.65 (m, 2H), 3.81 (dd, J=14.3, 3.0 Hz, 1H), 3.84 (s, 3H), 3.91–
4.03 (m, 1H), 3.93 (d, J=17.1 Hz, 1H), 4.07–4.15 (m, 2H), 4.21–4.27
(m, 1H), 4.32 (d, J=17.0 Hz, 1H), 6.45 (d, J=2.2 Hz, 1H), 6.53 (dd,
J=8.7, 2.0 Hz, 1H), 7.18 (dd, J=8.1, 5.9 Hz, 1H), 7.23 (d, J=7.3 Hz,
2H), 7.28 (td, J=7.6, 1.7 Hz, 2H), 7.57 ppm (t, J=8.4 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=22.7, 28.8, 30.4, 44.1, 45.2, 50.3, 54.4,
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1390 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
201.5 ppm; HRMS (ESI): calcd for C24H26N2NaO4 m/z 429.1785
[M+Na]+, found m/z 429.1788 [M+Na]+.
2-(2-(2-Bromoethoxy)phenyl)-2-phenethyl-1,3-dioxolane (13a):
Preparation of the ketal from 1-(2-(2-bromoethoxy)phenyl)-3-phe-
nylpropan-1-one like for 1a yielded a yellow oil (2.79 g, 99%);
1H NMR (500 MHz, CDCl3): d=2.48–2.55 (m, 2H), 2.61–2.68 (m, 2H),
3.68 (t, J=6.3 Hz, 2H), 3.83–3.92 (m, 2H), 4.02–4.11 (m, 2H), 4.33 (t,
J=6.5 Hz, 2H), 6.89 (d, J=8.2 Hz, 1H), 6.97 (t, J=7.6 Hz, 1H), 7.13
(t, J=7.4 Hz, 1H), 7.16 (d, J=7.6 Hz, 2H), 7.23 (t, J=7.4 Hz, 2H),
7.27 (t, J=7.3 Hz, 1H), 7.54 ppm (d, J=7.6 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=29.1, 30.1, 39.6, 64.8 (2C), 68.8, 109.9, 113.6,
121.0, 125.5, 127.7, 128.2 (2C), 128.4 (2C), 129.5, 130.3, 142.4,
155.7 ppm; HRMS (ESI): calcd for C19H22BrO3 m/z 377.0747 [M+H]+,
found m/z 377.0747 [M+H]+.
2-(2-(5-Methoxy-2-(3-phenylpropanoyl)phenoxy)ethyl)-hexahy-
dropyrrolo[1,2-a]pyrazine-1,4-dione (11b): Preparation like 6b
yielded a brown solid—mixture of partially deprotected tert-butyl
2-((2-(5-methoxy-2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)-ethyl)-
carbamoyl)pyrrolidine-1-carboxylate (96 mg, yield given for two
steps below); HRMS (ESI): calcd for C27H32N2NaO6 m/z 503.2153
[M+Na]+, found m/z 503.2155 [M+Na]+.
Deprotection like 6b yielded a yellowish semisolid (66 mg, 25%, 2
1
steps); H NMR (500 MHz, CDCl3): d=1.82–1.94 (m, 1H), 1.96–2.09
(m, 2H), 2.29–2.38 (m, 1H), 3.01 (t, J=8.1 Hz, 2H), 3.19–3.27 (m,
2H), 3.47–3.61 (m, 3H), 3.83 (s, 3H), 3.86 (d, J=16.4 Hz, 1H), 3.92
(t, J=7.9 Hz, 1H), 3.96–4.03 (m, 1H), 4.08–4.17 (m, 1H), 4.23 (d, J=
17.0 Hz, 1H), 4.22–4.28 (m, 1H), 6.43 (d, J=2.3 Hz, 1H), 6.53 (dd,
J=8.9, 2.2 Hz, 1H), 7.17 (t, J=7.1 Hz, 1H), 7.23 (d, J=7.3 Hz, 2H),
7.28 (t, J=7.6 Hz, 2H), 7.72 ppm (d, J=8.8 Hz, 1H); 13C NMR
(125 MHz, CDCl3): d=22.6, 28.6, 30.4, 44.9, 45.2, 46.1, 53.4, 55.6,
58.9, 65.8, 99.3, 105.8, 121.4, 125.9, 128.4 (2C), 128.5 (2C), 132.6,
159.0, 162.9, 164.1, 167.8, 198.9 ppm; HRMS (ESI): calcd for
C25H28N2NaO5 m/z 459.1890 [M+Na]+, found m/z 459.1894
[M+Na]+.
2,2,2-Trifluoro-N-(2-(2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)-
ethyl)-acetamide (14a): The bromide 13a (1 g, 1 eq, 2.65 mmol),
trifluoroacetamide (599 mg, 2 eq, 5.30 mmol), and tetrabutylammo-
nium bromide (85 mg, 0.1 eq, 0.27 mmol) were dissolved in dime-
thylformamide (DMF, 5 mL), and K2CO3 (733 mg, 2 eq, 5.3 mmol)
was added. The mixture was heated at 808C for 4.5 h, then TLC in-
dicated completion, and it was filtrated and evaporated. The crude
product was purified by column chromatography (petrol ether:E-
1
tOAc=30:1 to 4:1) to yield a colorless oil (708 mg, 65%); H NMR
(500 MHz, CDCl3): d=2.42–2.50 (m, 2H), 2.62–2.69 (m, 2H), 3.76
(dd, J=9.8, 5.0 Hz, 2H), 3.83–3.94 (m, 2H), 4.07–4.13 (m, 2H), 4.16
(t, J=5.0 Hz, 2H), 6.93 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H),
7.12–7.19 (m, 3H), 7.25 (t, J=7.4 Hz, 2H), 7.31 (t, J=7.7 Hz, 1H),
7.52 (d, J=7.9 Hz, 1H), 8.31 ppm (brs, 1H); 13C NMR (125 MHz,
CDCl3): d=29.9, 39.6, 39.9, 64.8 (2C), 67.1, 110.3, 114.3, 116.0 (q, J=
287.7 Hz), 121.6, 125.6, 127.1, 128.2 (2C), 128.3 (2C), 129.7, 130.4,
142.0, 155.7, 157.4 ppm (q, J=36.9 Hz); HRMS (ESI): calcd for
C21H22F3NNaO4 m/z 432.1393 [M+Na]+, found m/z 432.1397
[M+Na]+.
2-(2-(2-Acetylphenoxy)ethyl)hexahydropyrrolo[1,2-a]pyrazine-
1,4-dione (11c): The amine 9c (200 mg, 1 eq, 0.9 mmol), methyl
(2-chloroacetyl)prolinate (184 mg, 1 eq, 0.9 mmol), and Et3N
(0.16 mL, 1.3 eq, 1.2 mmol) were dissolved in 2-ethoxyethanol
(6 mL) and heated at reflux for 24 h. The solvent was evaporated,
and the residue was redissolved in a mixture of CH2Cl2 and toluene
(2:1, 100 mL) and extracted once with water (50 mL). The aqueous
phase was extracted CH2Cl2 (3”50 mL). The combined organic
phases were washed once with brine (70 mL), dried over Na2SO4,
and the solvent was evaporated. The product was purified by
column chromatography (CH2Cl2 to CH2Cl2:MeOH=20:1) twice to
yield a yellow semisolid (65 mg, 20%).
2-(2-(2-Phenethyl-1,3-dioxolan-2-yl)phenoxy)ethan-1-amine (9a),
Method B: The trifluoroacetamide derivative 14a (400 mg, 1 eq,
0.98 mmol) was dissolved in MeOH (2 mL) and water (2 mL), and
KOH (110 mg, 2 eq, 1.95 mmol) was added and stirred at RT for
2 h. Then the reaction mixture was concentrated and extracted
with EtOAc (3”50 mL). The combined organic extracts were dried
over Na2SO4 and evaporated to yield 9a as a colorless oil (305 mg,
99%).
2-(2-(2-(2-methyl-1,3-dioxolan-2-yl)phenoxy)ethyl)hexahydropyrro-
lo[1,2-a]pyrazine-1,4-dione: 1H NMR (500 MHz, CDCl3): d=1.73 (s,
3H), 1.85–1.97 (m, 1H), 1.99–2.12 (m, 2H), 2.36–2.44 (m, 1H), 3.51–
3.58 (m, 1H), 3.59–3.67 (m, 2H), 3.77–3.85 (m, 2H), 4.00–4.04 (m,
2H), 4.04–4.08 (m, 1H), 4.09–4.13 (m, 1H), 4.13–4.17 (m, 1H), 4.20
(d, J=17.0 Hz, 1H), 4.22 À4.28 (m, 1H), 4.55 (dd, J=17.0, 1.9 Hz,
1H), 6.87 (dd, J=8.2, 0.6 Hz, 1H), 6.94 (td, J=7.5, 1.1 Hz, 1H), 7.26
(ddd, J=8.1, 7.4, 1.8 Hz, 1H), 7.51 ppm (dd, J=7.6, 1.8 Hz);
13C NMR (125 MHz, CDCl3): d=22.6, 25.6, 28.8, 45.2, 46.6, 53.8, 59.0,
64.5 (2C), 66.4, 108.3, 112.8, 120.8, 126.8, 129.4, 130.8, 156.7, 163.7,
167.6 ppm; HRMS (ESI): calcd for C19H24N2NaO5 m/z 383.1577
[M+Na]+, found m/z 383.1583 [M+Na]+.
1-(2-(2-Hydroxy-3-(isopropylamino)propoxy)phenyl)-3-phenyl-
propan-1-one (16a): 1-(2-(Oxiran-2-ylmethoxy)phenyl)-3-phenyl-
propan-1-one (514 mg, 1.8 mmol) was dissolved in iPrNH2 (10 mL)
and heated for 21 h at reflux. The mixture was first concentrated
and then redissolved in EtOAc (10 mL) and acidified with 2 n HCl
(20 mL), which lead to the formation of a precipitate that was sep-
arated. The phases of the mother liquor were separated, and the
aqueous phase was basified with 2n ammonia (20 mL) and ex-
tracted with EtOAc (3”50 mL). The combined extracts were dried
over sodium sulfate and evaporated giving a first fraction of prod-
uct. The formed precipitate was separated, dissolved in CH2Cl2
(100 mL) and extracted with 2n ammonia (50 mL). The organic
phase was dried over Na2SO4 and evaporated, giving a second frac-
The ketal (35 mg, 1 eq, 0.09 mmol) was dissolved in CH3CN (2 mL);
aq. HCl solution (2n, 2 mL) was added, and it was stirred for 4 h at
RT. Then it was diluted with water and extracted with EtOAc (4”
40 mL), dried over Na2SO4, and evaporated to yield a colorless solid
(30 mg, 97%); 1H NMR (500 MHz, CDCl3): d=1.85–1.97 (m, 1H),
1.99–2.12 (m, 2H), 2.35–2.43 (m, 1H), 2.59 (s, 3H), 3.51–3.64 (m,
2H), 3.68–3.76 (m, 1H), 3.99 (d, J=16.7 Hz, 1H), 3.98–4.05 (m, 1H),
4.10 (t, J=7.6 Hz, 1H), 4.16–4.23 (m, 1H), 4.25–4.30 (m, 1H), 4.33
(d, J=16.7 Hz, 1H), 6.94 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.4 Hz, 1H),
7.44 (ddd, J=7.7, 7.4, 1.8 Hz, 1H), 7.67 ppm (dd, J=7.7, 1.7 Hz,
1H); 13C NMR (125 MHz, CDCl3): d=22.6, 28.7, 31.4, 45.2, 46.1, 53.3,
59.0, 65.8, 112.5, 121.2, 130.3, 133.4, 128.8, 157.0, 163.1, 167.8,
199.6 ppm; HRMS (ESI): calcd for C17H20N2NaO4 m/z 339.1315
[M+Na]+, found m/z 339.1321 [M+Na]+.
1
tion of product as a yellow oil (combined 353 mg, 57%); H NMR
(500 MHz, CDCl3): d=1.05 (d, J=6.3 Hz, 6H), 2.67 (dd, J=12.1,
7.7 Hz, 1H), 2.74 (septet, J=6.3 Hz, 1H) 2.83 (dd, J=12.0, 3.8 Hz,
1H), 3.03 (t, J=7.7 Hz, 2H), 3.33 (td, J=7.9, 1.8 Hz, 2H), 3.95–4.01
(m, 1H), 4.07 (d, J=5.0 Hz, 2H), 6.96 (d, J=8.5 Hz, 1H), 7.01 (t, J=
7.6 Hz, 1H), 7.18 (t, J=7.3 Hz, 1H), 7.23 (d, J=7.6 Hz, 2H), 7.28 (t,
J=7.6 Hz, 2H), 7.44 (t, J=7.9 Hz, 1H), 7.66 ppm (d, J=7.6 Hz, 1H);
13C NMR (125 MHz, CDCl3): d=22.8/23.0, 30.2, 45.1, 48.9, 49.3, 68.1,
71.4, 113.0, 121.0, 125.9, 128.3 (2C), 128.4 (2C), 130.2, 133.4, 141.5,
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1391 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
157.6, 201.5 ppm; HRMS (ESI): calcd for C21H28NO3 m/z 342.2064
[M+H]+, found m/z 342.2070 [M+H]+.
13.6/14.0 (t, J=4.6 Hz/t, J=4.6 Hz), 29.9/30.0, 40.0, 49.5/49.7, 55.2/
55.4 (t, J=21.6 Hz/t, J=23.1 Hz), 64.7/64.8, 69.1/69.3, 72.6, 110.3,
114.4, 118.0 (t, J=243.8 Hz), 121.0, 125.6, 127.1, 128.3 (4C), 129.6,
130.1, 142.1, 156.6 ppm; 19F NMR (470 MHz, CDCl3): d=(Diastereo-
mer 1) À124,4 (ddd, J=280.7, 56.8, 10.0 Hz, 1F), À127.9 (ddd, J=
281.0, 56.7, 14.0 Hz, 1F), (Diastereomer 2) À125.0 (ddd, J=281.2,
56.5, 10.3 Hz, 1F), À127.3 (ddd, J=281.0, 56.5, 12.8 Hz, 1F).
Deprotection like for 13b yielded a yellow oil (232 mg, 84%);
1H NMR (500 MHz, CDCl3): d=1.13 (d, J=6.5 Hz, 3H), 2.71–2.98 (m,
3H), 3.03 (t, J=7.7 Hz, 2H), 3.23–3.36 (m, 2H), 3.92–4.00 (m, 1H),
4.01–4.15 (m, 2H), 5.59 (td, J=56.5, 3.8 Hz, 1H)/5.60 (td, J=56.5,
3.8 Hz, 1H), 6.96 (d, J=8.2 Hz, 1H), 7.02 (t, J=7.6 Hz, 1H), 7.20 (t,
J=6.9 Hz, 1H), 7.23 (d, J=7.6 Hz, 2H), 7.28 (t, J=7.6 Hz, 2H),
7.44 ppm (t, J=7.9 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=13.8/
14.0 (t, J=3.9 Hz/t, J=3.9 Hz), 30.2, 45.0, 49.6, 55.1 (t, J=22.4 Hz),
68.4/68.6, 71.3/71.4, 113.3, 117.6 (t, J=244.3 Hz), 121.1, 126.0, 128.3
(2C), 128.4 (2C), 130.0, 130.1, 133.4, 141.5, 157.6, 201.4 ppm;
19F NMR (470 MHz, CDCl3): d=(Diastereomer 1) À125.6 (ddd, J=
281.3, 56.7, 10.6 Hz, 1F), À127.6 (ddd, J=281.4, 56.4, 13.4 Hz, 1F),
(Diastereomer 2) À125,6 (ddd, J=281.3, 56.7, 10.6 Hz, 1F), À127.5
(ddd, J=281.4, 56.4, 13.1 Hz, 1F); HRMS (ESI): calcd for C21H26F2NO3
m/z 378.1875 [M+H]+, found m/z 378.1878 [M+H]+.
The hydrochloride was prepared like for 5a and collected by filtra-
tion to yield a white solid (181 mg, 83%); HRMS (ESI): calcd for
C21H28NO3 m/z 342.2064 [M+H]+, found m/z 342.2068 [M+H]+.
1-(2-(3-((1-Fluoropropan-2-yl)amino)-2-hydroxypropoxy)phenyl)-
3-phenylpropan-1-one (16b): The ketal 4a (501 mg, 1 eq,
1.46 mmol) and NaBH3CN (368 mg, 4 eq, 5.86 mmol) were dis-
solved in MeOH (10 mL), and fluoroacetone (357 mg, 3 eq,
4.69 mmol) and acetic acid (0.2 mL) were added. The mixture was
heated for 4 h at reflux, then it was diluted with CH2Cl2 (40 mL)
and with 2 n NaHCO3 solution (40 mL). The organic phase was
dried over Na2SO4, filtered, and concentrated to yield a yellow oil
(596 mg, quant., mixture of diasteromers) that was used without
further purification. 1-((1-fluoropropan-2-yl)amino)-3-(2-(2-pheneth-
1
yl-1,3-dioxolan-2-yl)phenoxy)-propan-2-ol H NMR (500 MHz, CDCl3):
d=1.09 (t, J=7.4 Hz, 3H), 2.44 (t, J=8.8 Hz, 2H), 2.57–2.72 (m,
2H), 2.79–2.91 (m, 2H), 2.94–3.05 (m, 1H), 3.81–3.93 (m, 2H), 3.94–
4.01 (m, 1H), 4.01–4.15 (m, 3H), 4.15–4.46 (m, 3H), 6.90 (d, J=
8.2 Hz, 1H), 6.96 (t, J=7.6 Hz, 1H), 7.09–7.16 (m, 1H), 7.15 (d, J=
7.6 Hz, 2H), 7.23 (t, J=7.3 Hz, 2H), 7.28 (t, J=7.5 Hz, 1H), 7.48 ppm
3
(d, J=7.6 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=16.2 (d, J13C-19F=
2
7.7 Hz), 29.9, 39.9, 49.4, 53.1(d, J13C-19F=18.4 Hz), 64.7, 64.8, 69.2,
Hydrochloride formation like for 13a yielded
a white solid
72.7, 87.0/87.1 (d, 1J13C-19F=168.5 Hz), 110.3, 114.4, 120.9, 125.6,
127.1, 128.2 (2C), 128.3 (2C), 129.6, 130.1, 142.1, 156.7 ppm;
19F NMR (470 MHz, CDCl3): d=À224.5 (dt, J=47.3, 47.3, 16.9 Hz,
1F), À224.7 (dt, J=47.5, 47.5, 17.2 Hz, 1F).
(123 mg, 53%). HRMS (ESI): calcd for C21H26F2NO3 m/z 378.1875
[M+H]+, found m/z 378.1882 [M+H]+.
In silico modeling
The product from the last step (596 mg, 1 eq, 1.48 mmol) was dis-
solved in EtOAc (50 mL), and aq. HCl (2n, 50 mL) was added and
stirred for 1 h. Then the phases were separated, and the aqueous
phase was extracted with EtOAc (2”60 mL). The combined organic
phases were washed once with 2n NaHCO3 solution (50 mL), dried
over Na2SO4, and concentrated to a yellow oil (451 mg, 86%, mix-
ture of diasteromers); 1H NMR (500 MHz, CDCl3): d=1.10 (t, J=
7.1 Hz, 3H), 2.79 (dd, J=12.0, 7.6 Hz, 1H), 2.88–2.99 (m, 2H), 3.03
(t, J=7.7 Hz, 2H), 3.24–3.38 (m, 2H), 3.99–4.14 (m, 3H), 4.20–4.26
(m, 0.5H), 4.29–4.35 (m, 1H), 4.38–4.44 (m, 0.5H), 6.97 (d, J=
8.5 Hz, 1H), 7.02 (t, J=7.3 Hz, 1H), 7.19 (t, J=7.3 Hz, 1H), 7.23 (d,
J=7.3 Hz, 2H), 7.28 (t, J=7.6 Hz, 2H), 7.45 (t, J=7.7 Hz, 1H),
7.67 ppm (d, J=7.9 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=15.9/
16.1 (d, J(13C-19F)=6.1/7.7 Hz), 30.2, 44.9, 49.2, 53.1/53.2 (d, J(13C-
19F)=18.4 Hz), 68.0, 71.2/71.4, 86.4/86.5 (d, 1J(13C-19F)=170.1/
170.2 Hz), 113.1/113.2, 121.1, 126.0, 128.2/128.3, 128.4 (4C), 130.3,
133.5, 141.4, 157.6, 201.5 ppm; 19F NMR (470 MHz, CDCl3): d=
À225.2 (dt, J=47.5, 47.5, 17.7 Hz, 1F), À225.3 (dt, J=47.0, 47.0,
17.7 Hz, 1F); HRMS (ESI): calcd for C21H27FNO3 m/z 360.1969
[M+H]+, found m/z 360.1974 [M+H]+.
BCRP model: Data for building the BCRP inhibition prediction
model was taken as-is from Ref. [14a]. The dataset contains 433 in-
hibitors and 545 noninhibitors. Model building is described in
Ref. [27]. Briefly, extended connectivity fingerprints (ECFP) were
computed for each molecule of the training set using RDKit,[28]
with a bit vector length of 1024 and a radius of 4. Logistic regres-
sion was built in Python with the scikit-learn library.[29] The penalty
and C parameters were optimized with GridSearch.
P-gp model: Data for building the P-gp inhibition prediction
model was taken from Ref. [14b]. It was curated using the proce-
dure described in,[30] and which can be summarized as follows: in-
organic compounds were removed, salts were removed from mix-
tures, compounds containing rare atoms were removed, chemo-
types were normalized using the clean 2D, aromatize, mesomerize,
neutralize, tautomerize, and all transform options, duplicated com-
pounds were removed, and permanently charged compounds
were removed. MOE 2011.10 and ChemAxon 5.11.3 software were
used. The cleaned dataset contains 627 inhibitors and 553 non-
inhibitors of P-gp. ECFP fingerprints were computed for each mole-
cule of the training set using RDKit,[28] with a bit vector length of
1024 and a radius of 4. Support Vector Machine (SVM) with RBF
kernel was built in Python with the scikit-learn library.[16] The
gamma and C parameters were optimized with GridSearch.
3
2
Hydrochloride formation like for 13a yielded a white solid (84 mg,
70%); HRMS (ESI): calcd for C21H27FNO3 m/z 360.1969 [M+H]+,
found m/z 360.1969 [M+H]+.
1-(2-(3-((1,1-Difluoropropan-2-yl)amino)-2-hydroxypropoxy)-
phenyl)-3-phenylpropan-1-one(16c): Reductive amination using
1,1-difluoroacetone was performed like for 16b yielding a yellow
oil (313 mg, quant., mixture of diastereomers), 1-((1,1-Difluoropro-
pan-2-yl)amino)-3-(2-(2-phenethyl-1,3-dioxolan-2-yl)phenoxy)pro-
Model validation: Both models were validated by 10-fold cross-
validation. Traditional confusion matrix-related statistics are report-
ed, as well as the area under the receiver operating characteristic
(ROC) curve (AUC) that specifically evaluate the ranking capability
of the models.
1
pan-2-ol H NMR (500 MHz, CDCl3): d=1.16 (dd, J=6.3, 4.4 Hz, 3H),
2.38–2.48 (m, 2H), 2.56–2.72 (m, 2H), 2.80–3.01 (m, 3H), 3.82–3.92
(m, 2H), 3.92–4.02 (m, 1H), 4.02–4.14 (m, 3H), 4.16–4.24 (m, 1H),
5.63 (td, J=56.7, 4.0 Hz, 1H), 6.90 (d, J=8.5 Hz, 1H), 6.97 (t, J=
7.6 Hz, 1H), 7.09–7.18 (m, 3H), 7.19–7.25 (m, 2H), 7.28 (t, J=7.2 Hz,
1H), 7.48 ppm (d, J=7.9 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=
In vitro inhibition studies
Materials: RPMI 1640 media for cell culture was purchased from
LifeTechnologies (Rockville, MD, USA). Supplements for cell culture,
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1392 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
including fetal bovine serum (FBS), the antibiotics penicillin and
streptomycin, as well as vincristine for selection of CCRF VCR1000
cells were purchased from Sigma (St. Louis, MO, USA). The fluores-
cent substrates mitoxantrone and daunorubicin as well as verapa-
mil, Ko143, DMSO, PBS (phosphate buffered saline), and all com-
pounds used for HPMI buffer preparation were also purchased
from Sigma.
Top À Bottom
1 þ 10
Y ¼ Bottom þ
ð1Þ
log IC
ÀX ÁHillslope
Þ
ð
ð
50apparentÞ
To correct for the expression-level dependency of IC50apparent values
and the pump-leak kinetics as reviewed in W. Stein,[34] IC50apparent
values were calculated using Equation (2), where the fluorescence
intensity at zero inhibitor concentration [I]0 is according the
“Bottom” level and the fluorescence intensity at infinite inhibitor
Cell culture: The human myeloid leukemia PLB985 parental and
stably expressing BCRP cell lines[31] were kindly provided by B. Sar-
kadi (Institute of Enzymology, Research Centre for Natural Sciences,
Budapest, Hungary) and K. Nemet (Creative Cell Ltd., Budapest,
Hungary). Both cell lines were cultured in RPMI 1640 supplemented
with 10% FBS and 100 IUmLÀ1 of penicillin and streptomycin. The
human T-lymphoblast cell line CCRF VCR1000, overexpressing P-
gp, was obtained by stepwise selection of CCRF-CEM in vincristine-
containing medium[32] and was kindly provided by V. Gekeler
(Altana-Pharma formerly Byk-Gulden, Konstanz, Germany). CCRF
VCR1000 cells were cultured in RPMI 1640 medium containing
20% FBS and were treated regularly with vincristine (1 mgmLÀ1) for
about 72 h, followed by centrifugation (300 g, 5 min, RT) and resus-
pension in normal culture medium. All cell lines were maintained
at 378C in an atmosphere containing 5% CO2 with 95% relative
humidity.
concentration [I] is according to the “Top” level given by nonlin-
1
ear regression analysis.[35] At least three independent experiments
were performed for each compound showing an IC50 value less
than or equal to 10 mm.
fluorescence intensity at ½IŠ
0
ð2Þ
IC50 ¼ IC50apparent
Á
fluorescence intensity at ½IŠ
1
Acknowledgements
The authors thank Volker Gekeler for providing us with the CCRF
VCR1000 cells as well as Balµzs Sarkadi and Katalin Nemet for
providing PLB985 BCRP cells. The authors also thank Martin Zehl
for performing the high-resolution mass spectrometry measure-
ments. The authors are grateful for financial support provided by
the Austrian Science Fund (FWF) (SFB35, F3502, F3509, F3520) as
well as by the Hochschuljubiläumsstiftung (H-287748/2014). The
authors thank ChemAxon for kindly providing us with their soft-
ware. Furthermore, the authors acknowledge support through
the eTOX project (www.etoxproject.eu) from the Innovative Medi-
cines Initiative Joint Undertaking under Grant Agreement
n8 115002, resources of which are composed of financial contri-
bution from the European Union’s Seventh Framework Pro-
gramme (FP7/2007-2013) and European Federation of Pharma-
ceutical Industries and Associations (EFPIA) companies’ in-kind
contributions.
Steady-state accumulation experiments: For BCRP and P-gp inhibi-
tion studies, the steady-state accumulation of mitoxantrone (7 mm)
and daunorubicin (3 mm), respectively, was performed as previously
described.[33] Briefly, cells were harvested, pelleted (300 g, 5 min,
48C), diluted to a concentration of 12”106 cellsmLÀ1 in HPMI
(10 mm Hepes, 120 mm NaCl, 5 mm KCl, 0.4 mm MgCl2, 0.04 mm
CaCl2, 10 mm glucose, 10 mm NaHCO3, 5 mm Na2HPO4, pH 7.4 with
NaOH) and kept on ice for further processing. For each data point,
the cell suspension (25 mL) was preincubated for 5 min at 378C
with 2% DMSO (25 mL) in HPMI alone (DMSO control) or containing
different concentrations of test compound solutions. Thereafter,
cells were preloaded with daunorubicin (50 mL) or mitoxantrone
(50 mL) for 30 or 20 min at 378C, respectively, to reach a final con-
centration of 3 or 7 mm, respectively. The final DMSO concentration
did not exceed 1.5%. Preloading was stopped by chilling cells on
ice for 5 min followed by the addition of 400 mL ice-cold PBS. Cells
were pelleted (300 g, 5 min, 48C), the supernatant was discarded,
and the cell pellet was resuspended in ice-cold PBS (150 mL). Until
measurement, cells were kept on ice and in the dark. Immediately
before cell fluorescence measurement by flow cytometry (BD
Accuri C6 and BD FACSCalibur, Becton Dickinson, San Jose, CA,
USA; MACSQuant, MiltenyiBiotec GmbH, BergischGladbach, Germa-
ny), 50 mL of 4’,6-diamidin-2-phenylindole solution (DAPI, 4 mgmLÀ1
in PBS) was added to gate out dead cells. 100 mm verapamil and
1 mm Ko143 were used as positive controls for P-gp and BCRP in-
hibition, respectively.
Keywords: breast cancer resistance protein
· inhibitor ·
p-glycoprotein · polypharmacology · propafenone
[1] M. Dean, A. Rzhetsky, R. Allikmets, Genome Res. 2001, 11, 1156–1166.
[2] a) F. Thiebaut, T. Tsuruo, H. Hamada, M. M. Gottesman, I. Pastan, M. C.
Willingham, Proc. Natl. Acad. Sci. USA 1987, 84, 7735–7738; b) M. Malie-
paard, G. L. Scheffer, I. F. Faneyte, M. A. van Gastelen, A. C. Pijnenborg,
A. H. Schinkel, M. J. van De Vijver, R. J. Scheper, J. H. Schellens, Cancer
Res. 2001, 61, 3458–3464.
[3] J. Konig, F. Muller, M. F. Fromm, Pharmacol. Rev. 2013, 65, 944–966.
[4] a) K. S. Fenner, M. D. Troutman, S. Kempshall, J. A. Cook, J. A. Ware, D. A.
Smith, C. A. Lee, Clin. Pharmacol. Ther. Ser. 2009, 85, 173–181; b) A. J.
Sadeque, C. Wandel, H. He, S. Shah, A. J. Wood, Clin. Pharmacol. Ther.
Ser. 2000, 68, 231–237; c) W. Zhang, B. N. Yu, Y. J. He, L. Fan, Q. Li, Z. Q.
Liu, A. Wang, Y. L. Liu, Z. R. Tan, J. Fen, Y. F. Huang, H. H. Zhou, Clin.
Chim. Acta 2006, 373, 99–103; d) J. E. Keskitalo, M. K. Pasanen, P. J. Neu-
vonen, M. Niemi, Pharmacogenomics 2009, 10, 1617–1624.
[5] B. Sarkadi, L. Homolya, G. Szakacs, A. Varadi, Physiol. Rev. 2006, 86,
1179–1236.
IC50 value measurements and calculations: For compounds showing
an inhibitory effect of 50% inhibition compared with the positive
control (100 mm verapamil for P-gp and 1 mm Ko143 for BCRP) at
a concentration of 100 mm, IC50apparent values were estimated meas-
uring at least eight different compound concentrations and per-
forming nonlinear regression analyses (GraphPad Prism 6, “log(A-
gonist) vs. response—Variable slope ”), for which Equation (1) was
used, where X is the log of compound concentration, Y is the re-
sponse in fluorescent intensity units, “Bottom” and “Top” are the
lower and the higher plateaus of the nonlinear fit curve, respec-
tively, IC50apparent refers to the IC50apparent value, and HillSlope is
a factor that describes the steepness of the curve.
[6] G. Szakµcs, A. Vµradi, C. Özvegy-Laczka, B. Sarkadi, Drug Discovery Today
2008, 13, 379–393.
[7] T. A. Brooks, D. R. Kennedy, D. J. Gruol, I. Ojima, M. R. Baer, R. J. Bernacki,
Anticancer Res. 2004, 24, 409–415.
[8] a) N. A. Colabufo, F. Berardi, R. Perrone, S. Rapposelli, M. Digiacomo, A.
Balsamo, J. Med. Chem. 2006, 49, 6607–6613; b) N. A. Colabufo, F. Berar-
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1393 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
di, R. Perrone, S. Rapposelli, M. Digiacomo, M. Vanni, A. Balsamo, J. Med.
Chem. 2008, 51, 1415–1422.
Society National Meeting & Exposition, August 16–20, 2015, Boston
(USA)..
[9] H. M. Sim, K. Y. Loh, W. K. Yeo, C. Y. Lee, M. L. Go, ChemMedChem 2011,
6, 713–724.
[24] G. Ecker, M. Huber, D. Schmid, P. Chiba, Mol. Pharmacol. 1999, 56, 791–
796.
[10] M. Kühnle, M. Egger, C. Müller, A. Mahringer, G. Bernhardt, G. Fricker, B.
Kçnig, A. Buschauer, J. Med. Chem. 2009, 52, 1190–1197.
[11] G. Valdameri, C. Gauthier, R. Terreux, R. Kachadourian, B. J. Day, S. M. B.
Winnischofer, M. E. M. Rocha, V. Frachet, X. Ronot, A. Di Pietro, A. Bou-
mendjel, J. Med. Chem. 2012, 55, 3193–3200.
[12] G. Valdameri, E. Genoux-Bastide, B. Peres, C. Gauthier, J. Guitton, R. Ter-
reux, S. M. B. Winnischofer, M. E. M. Rocha, A. Boumendjel, A. Di Pietro,
J. Med. Chem. 2012, 55, 966–970.
[25] G. Ecker, C. R. Noe, W. Fleischhacker, Monatsh. Chem. 1997, 128, 53–59.
[26] T. Horaguchi, C. Tsukada, E. Hasegawa, T. Shimizu, T. Suzuki, K. Tane-
mura, J. Heterocycl. Chem. 1991, 28, 1261–1272.
[27] F. Sanz, P. Carrió, O. López, L. Capoferri, D. P. Kooi, N. P. E. Vermeulen,
D. P. Geerke, F. Montanari, G. F. Ecker, C. H. Schwab, T. Kleinçder, T.
Magdziarz, M. Pastor, Mol. Inf. 2015, 34, 477–484.
[28] RDKit: Open-Source Cheminformatics Software Version 2015.03.1;
www.rdkit.org.
[13] J. Cramer, S. Kopp, S. E. Bates, P. Chiba, G. F. Ecker, ChemMedChem 2007,
2, 1783–1788.
[14] a) F. Montanari, G. F. Ecker, Mol. Inf. 2014, 33, 322–331; b) F. Broccatelli,
E. Carosati, A. Neri, M. Frosini, L. Goracci, T. I. Oprea, G. Cruciani, J. Med.
Chem. 2011, 54, 1740–1751.
[15] a) R. Devakaram, D. S. C. Black, K. T. Andrews, G. M. Fisher, R. A. Davis, N.
Kumar, Bioorg. Med. Chem. 2011, 19, 5199–5206; b) G. Sagrera, G.
Seoane, Synthesis 2010, 2776–2786.
[29] F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel,
M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A.
Passos, D. Cournapeau, M. Brucher, M. Perrot, É. Duchesnay, J. Mach.
Learn. Res. 2011, 12, 2825–2830.
[30] M. Pinto, M. Trauner, G. F. Ecker, Mol. Inf. 2012, 31, 547–553.
[31] C. Özvegy-Laczka, G. Vµrady, G. Kçblçs, O. Ujhelly, J. Cervenak, J. D.
Schuetz, B. P. Sorrentino, G.-J. Koomen, A. Vµradi, K. NØmet, B. Sarkadi,
J. Biol. Chem. 2005, 280, 4219–4227.
[16] F. Ono, H. Takenaka, T. Fujikawa, M. Mori, T. Sato, Synthesis 2009, 1318–
1322.
[32] R. Boer, V. Gekeler, W. R. Ulrich, P. Zimmermann, W. Ise, A. Schoedl, S.
Haas, Eur. J. Cancer 1996, 32, 857–861.
[17] S. W. Laws, J. R. Scheerer, J. Org. Chem. 2013, 78, 2422–2429.
[18] J. B. Bremner, E. J. Browne, I. W. K. Gunawardana, Aust. J. Chem. 1984,
37, 129–141.
[33] L. Homolya, Z. Hollo, M. Muller, E. B. Mechetner, B. Sarkadi, Br. J. Cancer
1996, 73, 849–855.
[34] W. D. Stein, Physiol. Rev. 1997, 77, 545–590.
[19] K. V. Nikitin, N. P. Andryukhova, Can. J. Chem. 2004, 82, 571–578.
[20] D. Landini, M. Penso, J. Org. Chem. 1991, 56, 420–423.
[21] P. Chiba, G. Ecker, D. Schmid, J. Drach, B. Tell, S. Goldenberg, V. Gekeler,
Mol. Pharmacol. 1996, 49, 1122–1130.
[35] T. Litman, T. Skovsgaard, W. D. Stein, J. Pharmacol. Exp. Ther. 2003, 307,
846–853.
[22] P. Chiba, D. Annibali, M. Hitzler, E. Richter, G. Ecker, Farmaco 1998, 53,
357–364.
[23] “Topliss batchwise scheme reviewed in the era of Open Data”, L. Richter,
G. F. Ecker, abstract #CINF 84, presented at the 250th American Chemical
Received: December 21, 2015
Revised: February 3, 2016
Published online on March 10, 2016
ChemMedChem 2016, 11, 1380 – 1394
www.chemmedchem.org
1394 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim