Unveiling the "three-Finger Pharmacophore" Required for p53-MDM2 Inhibition by Saturation-Transfer Difference (STD) NMR Initial Growth-Rates Approach

Article abstract of DOI:10.1002/chem.201600114

Inhibitors of the p53-MDM2 protein-protein interaction are emerging as a new and validated approach to treating cancer. Herein, we describe the synthesis and inhibitory evaluation of a series of isoquinolin-1-one analogues, and highlight the utility of an initial growth-rates saturation-transfer difference (STD) NMR approach supported by protein-ligand docking to investigate p53-MDM2 inhibition. The approach is illustrated by the study of compound 1, providing key insights into the binding mode of this kind of MDM2 ligands and, more importantly, readily unveiling the previously proposed three-finger pharmacophore requirement for p53-MDM2 inhibition. Molecular modelling: Inhibitors of the p53-MDM2 protein-protein interaction are emerging as a new and validated approach to treating cancer. Herein, the synthesis and inhibitory evaluation of a series of isoquinolin-1-one analogues is described, and the utility of an initial growth-rates STD NMR approach supported by protein-ligand docking to investigate p53-MDM2 inhibition is highlighted (see figure).

Full text of DOI:10.1002/chem.201600114

DOI: 10.1002/chem.201600114
Communication
&
NMR Spectroscopy
Unveiling the “Three-Finger Pharmacophore” Required for p53–
MDM2 Inhibition by Saturation-Transfer Difference (STD) NMR
Initial Growth-Rates Approach
Jesus Angulo,*[a] Sarah A. Goffin,[a] Daivik Gandhi,[a] Mark Searcey,[a, b] and Lesley A. Howell*[a]
iting the p53-MDM2 interaction with small molecules, repre-
Abstract: Inhibitors of the p53-MDM2 protein–protein in-
teraction are emerging as a new and validated approach
to treating cancer. Herein, we describe the synthesis and
inhibitory evaluation of a series of isoquinolin-1-one ana-
logues, and highlight the utility of an initial growth-rates
saturation-transfer difference (STD) NMR approach sup-
ported by protein–ligand docking to investigate p53-
MDM2 inhibition. The approach is illustrated by the study
of compound 1, providing key insights into the binding
mode of this kind of MDM2 ligands and, more important-
ly, readily unveiling the previously proposed three-finger
pharmacophore requirement for p53-MDM2 inhibition.
sents an attractive and viable approach to treating cancer.
NMR spectroscopy is one of the most powerful spectroscop-
ic techniques to study biomolecular interactions and has been
applied to the discovery of protein–protein inhibitors,[14] some
of them focused on the MDM2–p53 interaction.[15–17] In particu-
1
lar, Holak and co-workers recently devised a smart 2D H,15N
NMR-based method to test inhibition of PPIs, called AIDA, that
they also applied to the MDM2–p53 interaction.[18,19] However,
although that method provides relevant information about the
protein residues of MDM2 involved in binding, the information
about the bioactive conformation and the mode of binding of
the ligand is lost. This kind of structural information is attaina-
ble by saturation-transfer difference (STD) NMR spectrosco-
py,[20–22] if a thorough quantitative analysis of the experiments
is carried out by means of the STD build-up curves.[23–25]
Herein, we report the synthesis of a series of isoquinolin-1-
one analogues, their p53–MDM2 inhibitory activity as was de-
termined by using a fluorescence polarisation assay, and a STD
NMR initial growth-rates approach to identify the binding
mode of one of the lead compounds. Notably, the results sug-
gest that STD NMR is an easy and powerful tool to verify the
proposed three-finger pharmacophore requirement for p53-
MDM2 inhibition, providing in addition key structural informa-
tion regarding its interactions with the hydrophobic groove of
MDM2, for future lead optimisation.
Protein–protein interactions (PPIs) had been considered “un-
druggable” primarily due to large surface areas and their flat,
featureless and hydrophobic nature.[1–3] However, the success
of small-molecule inhibitors, such as the Nutlins,[4] p53-MDM2
inhibitors and Navitoclax,[5,6] a dual inhibitor of Bcl-2 and Bcl-xL
(both of which are currently in clinical trials), have defied this
view point. PPIs still pose a considerable challenge to the me-
dicinal chemistry community, but they are an attractive drug
target due to their ability to modulate outcomes within cells,
hence, allowing greater control than classical drug targets,
such as enzymes or receptors.[7,8] The most widely studied PPI
is the p53-MDM2 paradigm.[9] The tumour-suppressor protein
p53 is a key transcription factor involved in regulating the cell
cycle and apoptosis.[10] It is often referred to as the guardian of
the genome[11] and plays a crucial role in cancer; nearly all tu-
mours show a defect in the p53 gene itself or in other nega-
tive regulatory proteins, such as MDM2 or MDMX.[12] In cancer
cells with wt-p53, over expression of either MDM2 and/or
MDMX supresses p53 activity and disrupts the apoptosis path-
way.[13] Therefore, the restoration of the p53 pathway, by inhib-
Our previous work on p53–MDM2 inhibitors has involved
the natural product chlorofusin,[26,27] and more recently, the
identification of a new small-molecule inhibitor inspired by our
studies of this natural product.[28] Of late, we have been work-
ing on producing a library of isoquinolin-1-one analogues. Iso-
quinoline-1-ones have been shown to inhibit the p53-MDM2
PPI with low micromolar activity.[29] We sought to develop
a procedure by which a relatively large number of diverse
compounds could be synthesised quickly with minimum purifi-
cation to identify potent small-molecule inhibitors of the p53–
MDM2 interaction.
A modified Castagnoli reaction was employed to synthesise
the isoquinolin-1-ones.[30] The reaction involves the condensa-
tion of a Schiff base with an acid anhydride (Scheme 1). Briefly,
a stoichiometric amount of a benzyl amine and an aldehyde
were reacted together under anhydrous conditions in the pres-
ence of magnesium sulphate to form the corresponding imine.
The reactions reached completion in 2–4 h (as was shown by
TLC), after which the solutions were filtered. The filtrate was
then added to one equivalent of homophthalic anhydride, and
[a] Dr. J. Angulo, S. A. Goffin, D. Gandhi, Prof. M. Searcey, Dr. L. A. Howell
School of Pharmacy, University of East Anglia
Norwich Research Park, Norwich, NR4 7TJ (UK)
E-mail: j.angulo@uea.ac.uk
l.howell@uea.ac.uk
[b] Prof. M. Searcey
School of Chemistry, University of East Anglia
Norwich Research Park, Norwich, NR4 7TJ (UK)
Supporting information for this article can be found under http://
dx.doi.org/10.1002/chem.201600114.
Chem. Eur. J. 2016, 22, 5858 – 5862
5858
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 1. Synthesis of isoquinolin-1-ones.
the reaction mixture was left overnight at room temperature.
The formed precipitate was filtered, washed with hot ethyl
acetate, and dried in vacuo. Those compounds that did not
precipitate were evaporated and purified by flash column
chromatography. A total of 60 compounds were prepared in
this manner (see the Supporting Information). The reaction can
produce either the cis or trans isomers or a mixture of both
with the major diastereoisomer expected to be the more ther-
modynamically stable trans product.[31]
Table 1. Binding of compounds 1–7 to MDM2 using the fluorescence po-
larisation assay.
Compd
R1
R2
IC50 [mm][a]
Ki [mm][b]
1
2
3
4
5
6
7
4-ClPh
4-BrPh
4-IPh
4-ClPh
4-BrPh
4-ClPh
4-BrPh
4-F
4-F
4-F
4-Cl
4-Cl
4-Br
4-Br
56.6
19.8
57.7
61.2
21.3
27.1
6.6
7.13
2.49
7.29
7.71
2.68
2.36
0.83
We had originally assumed the reaction would proceed race-
mically and initially proposed to evaluate the compounds as
racemic mixtures. However, upon further investigation, we
found the reaction did indeed exhibit some stereocontrol, be-
cause the compounds were found to be optically active. Look-
ing into the literature, there is some precedent for stereocon-
trol for this class of reaction: the condensation of o-anisylide-
nemethylamine with glutaric anhydride gave the trans piperi-
done product as the major diastereoisomer with the aromatic
ring occupying the axial position essentially fixing the stereo-
chemistry at this carbon.[32] This is a result of the planar nature
of the amide bond. Presumably, this is further enhanced in our
molecules due to the additional planar benzene fused to the
piperidone resulting in only one enantiomer formed as the
major product. An alternative explanation for the observation
of only one enantiomer is preferential crystallisation, also
known as resolution by entrainment, whereby the resolution is
performed without the use of a resolving agent.[33]
[a] Concentration of substrate required to decrease polarization fluores-
cence by 50%. Experiments were performed in triplicate. [b] Apparent in-
hibition constant. Experiments were performed in triplicate.
exchange conditions (Ki values in Table 1 suggest these mole-
cules might fall at the limits of the technique). Within the
weakest binders (1, 3 and 4), we selected 1, because it was the
most potent inhibitor (lowest Ki). The binding of 1 to human
MDM2 in solution was confirmed by STD NMR (Figure 1).
Strong saturation-transfer signals were observed in the differ-
ence spectrum (Figure 1, bottom). Besides providing a proof of
binding, structural information was obtained by a complete ki-
netics study of the evolution of STD intensities with the satura-
tion time of the protein. Briefly, the STD intensities were deter-
mined at different saturation times (STD build-up curves), and
the initial growth rates of each curve were obtained by mathe-
matical fitting; those initial slope values were then used to
map out the ligand epitope (see the Supporting Informa-
tion).[24] The presence of aromatic rings in combination with sa-
turated residues (e.g., aliphatics) in most of the structures of
protein–protein inhibitors (similar to the case of 1) makes this
approach very suitable, because “aromatic ring–protein” con-
tacts can be overstated, if a single large saturation time is em-
ployed, instead of a whole build-up analysis.[34]
The compounds were screened by using a fluorescence po-
larisation (FP) assay described previously.[28] Human MDM2 pro-
tein (17–125) was used in the polarisation assay and the wild-
type p53 peptide (residues 15–27) was used as a positive con-
trol and had an half maximal inhibitory concentration (IC50) of
14.45 mm and Ki of 1.82 mm. Screening the isoquinolin-1-ones
at 100 mm concentration revealed seven hits, which were fur-
ther evaluated over a wider concentration range to determine
their IC50’s (Table 1). Interestingly, all seven compounds were
found to be the trans isomer, and all were substituted with hal-
ogens indicating the importance of halides for MDM2 binding
ligands. Compound 7 was the most active with low micromo-
lar activity with compounds 2, 5 and 6, exhibiting similar activ-
ity to wild-type p53.
The initial growth rates of the curves were determined by
using a monoexponential model (see the Supporting Informa-
tion), and their relative distribution among the protons of
1 was calculated to map out the binding epitope of the ligand
for binding to human MDM2 (normalised STD values in
Figure 2). The binding epitope of 1 revealed significant struc-
tural information about the molecular recognition of this kind
of compound by human MDM2 (Figure 2). The three aromatic
residues of 1 constitute the main spatial contacts with the pro-
tein in the bound state, with the chloride substituted phenyl
ring (R1 in Scheme 1) showing the closest contacts. In contrast,
We chose compound 1 as our model compound for the
NMR binding study. Due to the structural similarity of 1–7,
comparable binding modes for all could be foreseen, and we
then decided to test one of the weakest ligands, because STD
NMR requires binding kinetics falling within the so-called fast-
Chem. Eur. J. 2016, 22, 5858 – 5862
www.chemeurj.org
5859
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
p53 binding site of human MDM2. The most polar part of
1 (the central six-membered ring and the methylene bridge)
makes fewer contacts with the protein and is accordingly more
solvent exposed in the bound state. This mode of binding is
reminiscent of that of some previously described MDM2–p53
inhibitors, strongly supporting that the initial growth-rates STD
NMR approach can be used as a simple technique to unveil
the so-called “three-finger pharmacophore” requirement for
MDM2–p53 inhibition (Figure 2).[9]
To get a 3D molecular model of the interaction of 1 with
human MDM2, we carried out docking calculations by using
the program Glide.[35,36] We used the published Cartesian coor-
dinates of the protein (PDB entry 1T4E) in complex with a ben-
zodiazepinedione.[37] The original ligand was removed, and
docking calculations were run with ligand 1. Although we
knew the ligand sample (without MDM2) had optical activity,
by using NMR spectroscopy, it was not possible to elucidate,
which enantiomer was in excess. For that reason, we carried
out the docking calculations with both enantiomers. Interest-
ingly, the 3R,4R-enantiomer led to the energetically most fa-
vourable docking solution, which in addition was in excellent
agreement with the experimental STD NMR data (Figure 3). For
the 3S,4S enantiomer, the best scored docking solution was
much higher in energy (above 17 kcalmolÀ1), and gave very
poor agreement with the experimental STD NMR data (see the
Supporting Information). In this way, the combined protocol of
STD initial growth rates and docking calculations allowed us to
identify the 3R,4R-enantiomer as the most active for MDM2 in
solution. The best-scored docking pose for the 3R,4R-enantio-
mer was further energy minimized and the solution is shown
in Figure 3.
Figure 1. Expanded view of the aromatic region of the STD NMR (298 K,
800 MHz, 1 s saturation time) of a sample containing an excess of 1 (1 mm
concentration) over the protein MDM2 (20 mm). The top spectrum corre-
1
sponds to the equilibrium intensities of 1 (1D reference H NMR spectrum),
whereas the bottom one shows the difference spectrum, in which the inten-
sities corresponds to transfer of magnetization from the protein upon bind-
ing (most intense STD signals highlighted in the spectrum and on the chem-
ical formula of 1, inset at the top).
A comparison of Figures 2 and 3 demonstrates that the 3D
docking molecular model of 1 bound to human MDM2 agrees
remarkably well with the experimental NMR results in solution
state. The three aromatic residues of 1 make close contacts
Figure 2. Binding epitope of 1 to the protein MDM2 as obtained by STD
NMR. The values indicate normalized STD values for each proton of 1. The
highest values results from very close contacts of the ligand to the MDM2
surface in the bound state (red), whereas the smallest ones indicate regions
of 1 being solvent exposed.
the six-membered ring tethering the aromatic residues togeth-
er showed the lowest STD intensities, along with the methyl-
ene bridge protons linking the fluoride-substituted phenyl ring
(Figure 2, in blue). The differences in relative STD values, al-
though small, are significant, because they correspond to
a slow kinetics (low micromolar) binder, for the standards of
STD NMR, and for that reason the criterion to split STD intensi-
ties in strong and weak was raised up to a level of 80% rela-
tive STD (Figure 2). The firmness of the interpretation of the
binding epitope of 1 relies on the accuracy of the STD initial
growth-rates approach followed on this work (see the Support-
ing Information).
Figure 3. 3D molecular model of the complex of 1 with human MDM2.
(a) Bound conformation (best docking pose) of 1 in the p53-binding pocket
of MDM2. (b) Mesh representation of the MDM2 surface closest to non-polar
hydrogens of 1 in the bound state (those observable in STD NMR). The ener-
getically most favourable docking solution (enantiomer 3R,4R) qualitatively
gives an excellent agreement with the observed STD NMR intensities.
In solution, 1 binds human MDM2 mainly through the
apolar aromatic moieties, which is compatible with the largely
hydrophobic character of the amino acid residues lining the
Chem. Eur. J. 2016, 22, 5858 – 5862
www.chemeurj.org
5860
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
with the protein in the bound state, with a higher predomi-
nance of contacts with amino acids with hydrophobic side-
chains (see map of protein–ligand contacts in the Supporting
Information). The chlorine-substituted phenyl ring is buried in
an internal cavity between GLY58 and LEU54, explaining the
largest amount of saturation transferred from the protein in
the NMR spectra. The carboxylate group, as well as the methyl-
ene bridge, are pointing towards the solvent, which agrees
very well with the observed lower STD intensities.
herein will be a powerful and simple method for distinguishing
between both (three- or four-finger) pharmacophore modes of
binding for new generations of MDM2 ligands. We envisage an
increased use of the STD NMR initial growth rates approach to
the design of protein–protein inhibitors, to verify the pharma-
cophore, and to determine the structural requirements for mo-
lecular recognition, extremely valuable information for the im-
provement of the small-molecule candidates to inhibit PPIs.
The STD NMR initial growth-rates approach and the model-
ling study of 1 has provided very relevant structural informa-
tion about the molecular recognition of this series of ligands
by the human MDM2 receptor. In this way, further improve-
ment of human MDM2 ligands should exploit this information;
for example, modifications must be on those parts of the li-
gands making fewer contacts with the protein (carboxylate
and methylene bridge). Indeed, an analogue substituted on
the methylene bridge with a methyl alcohol inhibited the PPI
with an IC50 of 15.01 mm. This is also in agreement with the
previous study by Holak and co-workers, as they showed that
changing the carboxylate of similar compounds to an amide
spacer did not affect the affinity.[29] Interestingly, the compari-
son of the binding mode of 1 with the published structure of
the complex with the benzodiazepinedione ligand (see the
Supporting Information), further highlights the two pending
halogenated aromatic residues as being the most important el-
ements for molecular recognition. In the case of the benzodia-
zepinedione ligand,[37] the fused aromatic ring bears a bulky
iodine atom that pushes it farther from the protein surface
than in the case of 1, so that the fused aromatic ring of 1 falls
in a significantly shifted position, compared to that ligand,
which supports that the molecular recognition is not specific
for that moiety. Yet, the matching of the two halogenated aro-
matic rings of 1 with the published benzodiazepinedione
structure is excellent (see the Supporting Information).
Acknowledgements
We thank the School of Pharmacy (UEA) for a studentship for
S.A.G. and for start-up funds to J.A. and the Swansea EPSRC
Mass Spectrometry Service for high-resolution mass spectra.
This work was supported by the Welcome Trust (102201/Z/13/
Z) a vacation scholarship for D.G. We also thank G. Richard Ste-
phenson for helpful discussions.
Keywords: cancer · molecular modelling · NMR spectroscopy ·
p53–MDM2 · proteins · saturation-transfer difference NMR
spectroscopy
[1] S. Jones, J. M. Thornton, Proc. Natl. Acad. Sci. USA 1996, 93, 13–20.
[2] O. Keskin, A. Gursoy, B. Ma, R. Nussinov, Chem. Rev. 2008, 108, 1225–
1244.
[3] A. J. Wilson, Chem. Soc. Rev. 2009, 38, 3289–3300.
[4] L. T. Vassilev, B. T. Vu, B. Graves, D. Carvajal, F. Podlaski, Z. Filipovic, N.
Kong, U. Kammlott, C. Lukacs, C. Klein, N. Fotouhi, E. A. Liu, Science
2004, 303, 844–848.
[5] M. Bruncko, T. K. Oost, B. A. Belli, H. Ding, M. K. Joseph, A. Kunzer, D.
Martineau, W. J. McClellan, M. Mitten, S.-C. Ng, P. M. Nimmer, T. Olters-
dorf, C.-M. Park, A. M. Petros, A. R. Shoemaker, X. Song, X. Wang, M. D.
Wendt, H. Zhang, S. W. Fesik, S. H. Rosenberg, S. W. Elmore, J. Med.
Chem. 2007, 50, 641–662.
[6] C. Tse, A. R. Shoemaker, J. Adickes, M. G. Anderson, J. Chen, S. Jin, E. F.
Johnson, K. C. Marsh, M. J. Mitten, P. Nimmer, L. Roberts, S. K. Tahir, Y.
Xiao, X. Yang, H. Zhang, S. Fesik, S. H. Rosenberg, S. W. Elmore, Cancer
Res. 2008, 68, 3421–3428.
In conclusion, we have reported the synthesis of a library of
isoquinolin-1-ones as potential inhibitors of the p53-MDM2
protein–protein interaction. Seven compounds were identified
with IC50 values in the low micromolar range. One of the com-
pounds, 1, was explored further by using STD NMR to deter-
mine its binding epitope for the hydrophobic groove of
human MDM2. These studies have shown the power of the ini-
tial growth rates STD NMR approach applied to biologically rel-
evant PPIs. In the particular case of the p53-MDM2 interaction,
the approach has demonstrated to be very useful for verifying
the three-finger pharmacophore requirement of small mole-
cules for inhibition, without the need for isotopic labelling of
the protein, and using small sub-stoichiometric amounts of the
receptor. The analysis of the whole STD build-up curve (initial
growth rates) is mandatory, because typical PPI inhibitors con-
tain protons with different relaxation properties that could
lead to misinterpreted epitopes if a single “one-saturation
time” experiment were used. Recently, a new type of p53-
MDM2 inhibitor, based on 6-chloroindole scaffolds, has demon-
strated that the “plasticity” of the p53-binding site on MDM2
allows some small molecules to show an extended four-point
pharmacophore model;[38] the STD NMR approach followed
[7] V. Azzarito, K. Long, N. S. Murphy, A. J. Wilson, Nat. Chem. 2013, 5, 161–
173.
[8] J. A. Wells, C. L. McClendon, Nature 2007, 450, 1001–1009.
[9] K. Khoury, A. Dçmling, Curr. Pharm. Des. 2012, 18, 4668–4678.
[10] K. H. Vousden, X. Lu, Nat. Rev. Cancer 2002, 2, 594–604.
[11] D. P. Lane, Nature 1992, 358, 15–16.
[12] M. Hollstein, D. Sidransky, B. Vogelstein, C. C. Harris, Science 1991, 253,
49–53.
[13] M. Wade, Y.-C. Li, G. M. Wahl, Nat. Rev. Cancer 2013, 13, 83–96.
[14] E. Barile, M. Pellecchia, Chem. Rev. 2014, 114, 4749–4763.
[15] O. Schon, A. Friedler, M. Bycroft, S. M. V. Freund, A. R. Fersht, J. Mol. Biol.
2002, 323, 491–501.
[16] D. C. Fry, S. D. Emerson, S. Palme, B. T. Vu, C.-M. Liu, F. Podlaski, J.
Biomol. NMR 2004, 30, 163–173.
[17] C. Riedinger, M. E. Noble, D. J. Wright, F. Mulks, I. R. Hardcastle, J. A. En-
dicott, J. M. McDonnell, Chem. Biol. Drug Des. 2011, 77, 301–308.
[18] L. D’Silva, P. Ozdowy, M. Krajewski, U. Rothweiler, M. Singh, T. A. Holak,
J. Am. Chem. Soc. 2005, 127, 13220–13226.
[19] M. Krajewski, U. Rothweiler, L. D’Silva, S. Majumdar, C. Klein, T. A. Holak,
J. Med. Chem. 2007, 50, 4382–4387.
[20] J. Angulo, P. M. Nieto, Eur. Biophys. J. 2011, 40, 1357–1369.
[21] M. Mayer, B. Meyer, Angew. Chem. Int. Ed. 1999, 38, 1784–1788; Angew.
Chem. 1999, 111, 1902–1906.
[22] M. Mayer, B. Meyer, J. Am. Chem. Soc. 2001, 123, 6108–6117.
[23] J. Angulo, P. M. Enríquez-Navas, P. M. Nieto, Chem. Eur. J. 2010, 16,
7803–7812.
Chem. Eur. J. 2016, 22, 5858 – 5862
www.chemeurj.org
5861
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[24] M. Mayer, T. L. James, J. Am. Chem. Soc. 2004, 126, 4453–4460.
[25] J. C. MuÇoz-García, E. Chabrol, R. R. Viv›s, A. Thomas, J. L. de Paz, J.
Rojo, A. Imberty, F. Fieschi, P. M. Nieto, J. Angulo, J. Am. Chem. Soc.
2015, 137, 4100–4110.
[26] J. P. Malkinson, M. Zloh, M. Kadom, R. Errington, P. J. Smith, M. Searcey,
Org. Lett. 2003, 5, 5051–5054.
[27] E. C. Y. Woon, M. Arcieri, A. F. Wilderspin, J. P. Malkinson, M. Searcey, J.
Org. Chem. 2007, 72, 5146–5151.
[28] M. M. D. Cominetti, S. A. Goffin, E. Raffel, K. D. Turner, J. C. Ramoutar,
M. A. O’Connell, L. A. Howell, M. Searcey, Bioorg. Med. Chem. Lett. 2015,
25, 4878–4880.
[34] J. Yan, A. D. Kline, H. Mo, M. J. Shapiro, E. R. Zartler, J. Magn. Reson. San
Diego Calif. 1997 2003, 163, 270–276.
[35] R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic, D. T.
Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry, D. E. Shaw, P.
Francis, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739–1749.
[36] T. A. Halgren, R. B. Murphy, R. A. Friesner, H. S. Beard, L. L. Frye, W. T. Pol-
lard, J. L. Banks, J. Med. Chem. 2004, 47, 1750–1759.
[37] B. L. Grasberger, T. Lu, C. Schubert, D. J. Parks, T. E. Carver, H. K. Koblish,
M. D. Cummings, L. V. LaFrance, K. L. Milkiewicz, R. R. Calvo, D. Maguire,
J. Lattanze, C. F. Franks, S. Zhao, K. Ramachandren, G. R. Bylebyl, M.
Zhang, C. L. Manthey, E. C. Petrella, M. W. Pantoliano, I. C. Deckman, J. C.
Spurlino, A. C. Maroney, B. E. Tomczuk, C. J. Molloy, R. F. Bone, J. Med.
Chem. 2005, 48, 909–912.
[29] U. Rothweiler, A. Czarna, M. Krajewski, J. Ciombor, C. Kalinski, V. Khazak,
G. Ross, N. Skobeleva, L. Weber, T. A. Holak, ChemMedChem 2008, 3,
1118–1128.
[38] M. Bista, S. Wolf, K. Khoury, K. Kowalska, Y. Huang, E. Wrona, M. Arcinie-
ga, G. M. Popowicz, T. A. Holak, A. Dçmling, Structure 2013, 21, 2143–
2151.
[30] M. Cushman, J. Gentry, F. W. Dekow, J. Org. Chem. 1977, 42, 1111–1116.
[31] M. Cushman, T.-C. Choong, J. T. Valko, M. P. Koleck, J. Org. Chem. 1980,
45, 5067–5073.
[32] M. Cushman, N. Castagnoli, J. Org. Chem. 1973, 38, 440–448.
[33] A. Collet, M. J. Brienne, J. Jacques, Chem. Rev. 1980, 80, 215–230.
Received: January 11, 2016
Published online on March 7, 2016
Chem. Eur. J. 2016, 22, 5858 – 5862
www.chemeurj.org
5862
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim