Revealing the Macromolecular Targets of Fragment-Like Natural Products

Article abstract of DOI:10.1002/anie.201504241

Fragment-like natural products were identified as ligand-efficient chemical matter for hit-to-lead development and chemical-probe discovery. Relying on a computational method using a topological pharmacophore descriptor and a drug database, several macromolecular targets from distinct protein families were expeditiously retrieved for structurally unrelated chemotypes. The selected fragments feature structural dissimilarity to the reference compounds and suitable target affinity, and they offer opportunities for chemical optimization. Experimental confirmation of hitherto unknown macromolecular targets for the selected molecules corroborate the usefulness of the computational approach and suggests broad applicability to chemical biology and molecular medicine. Target acquired: Hitherto unknown macromolecular targets of the fragment-like natural products goitrin, isomacroin, and graveolinine were discovered through the use of a computational target-prediction tool tailored for natural products. The results suggest that such methods will find application in target discovery for natural products and could inspire the design of new chemical entities for chemical biology and molecular medicine.

Full text of DOI:10.1002/anie.201504241

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201504241
German Edition: DOI: 10.1002/ange.201504241
Target Identification
Revealing the Macromolecular Targets of Fragment-Like Natural
Products**
Tiago Rodrigues, Daniel Reker, Jens Kunze, Petra Schneider, and Gisbert Schneider*
[
7]
Abstract: Fragment-like natural products were identified as
ligand-efficient chemical matter for hit-to-lead development
and chemical-probe discovery. Relying on a computational
method using a topological pharmacophore descriptor and
a drug database, several macromolecular targets from distinct
protein families were expeditiously retrieved for structurally
unrelated chemotypes. The selected fragments feature struc-
tural dissimilarity to the reference compounds and suitable
target affinity, and they offer opportunities for chemical
optimization. Experimental confirmation of hitherto unknown
macromolecular targets for the selected molecules corroborate
the usefulness of the computational approach and suggests
broad applicability to chemical biology and molecular med-
icine.
objective drug design. We have recently developed a soft-
ware tool for predicting the macromolecular targets of
[
4a]
structurally complex NPs. Conversely, small fragment-like
chemical entities are generally recognized as better-suited
starting points for hit-to-lead discovery programs. In this
study, we set out to identify the targets of fragment-like NPs
(molecular weight < 300 gmol ) and probe the underlying
molecular basis of macromolecular recognition.
By using molecular representation in terms of topological
pharmacophore features (CATS2 descriptor), we trained
a self-organizing map (SOM)
(Dictionary of Natural Products, DNP), together with 12 661
[
8]
À1 [9]
[10]
[11]
on a set of 210 213 NPs
[12]
small molecules with known targets (COBRA database).
This allowed us to span the combined pharmacophore feature
space of both drugs and NPs. The software modeled
qualitative ligand–receptor relationships by using statistically
interpreted similarities between drugs and NPs that co-
clustered in this comprehensive space. We derived a stat-
istical significance estimate (p value) for each individual
target prediction (see the Supporting Information). In
a previous study, we demonstrated that this method achieves
receiver-operating characteristic area under the curve
N
atural products (NPs) are the focus of strong interest as
[
1]
chemical matter for interrogating biological systems. Acting
as starting points for drug and chemical-probe discovery, they
comprise biologically prevalidated architectures that allow
the exploration of drug-relevant chemical space not covered
[5a]
[
2]
by fully synthetic small molecules. Consequently, computa-
tional tools have been deployed for enriching collections of
screening compounds by gathering structural and physico-
(ROC AUC) values of 0.88 based on an extensive retrospec-
[
3]
[5a]
chemical features from NPs. However, their applicability is
tightly bound to factual knowledge of the macromolecular
counterparts (e.g., target receptors) of the parent NP. The on-
and off-target interactions remain largely unknown for the
majority of these molecules, thus hindering the widespread
and efficient exploration of NP chemical motifs as leads for
tive study.
ROC AUC values can be interpreted as
probabilities for the algorithm to rank a correct ligand–
target pair higher than an incorrect pair.
The SOM algorithm afforded at least one highly confident
target prediction (p < 1%) for 80 024 (38%) of all of the NPs
annotated in the DNP, with an average of 0.9 targets per
chemical entity. From the 64 650 fragment-like NPs in the
DNP, 34 239 (53%) were confidently predicted to interact
with at least one target (average of 1.4 targets per fragment-
like NP). Importantly, the algorithm captured known ligand–
target relationships for fragment-like NPs that were not part
of the training data (see the Supporting Information). With
these target-engagement predictions in hand, we selected
goitrin, isomacroin, and graveolinine (Figure 1) for experi-
mental validation, taking into account: 1) their fragment-like
nature; 2) the confidence of the predictions (p < 1%, i.e., top
ranking); 3) their profound structural dissimilarity to the most
similar reference compound from the drug database (struc-
tural Tanimoto similarity ꢀ 0.2; Table 1), thus ensuring that
straightforward similarity searches would not have recog-
nized these potential ligand–receptor associations; 4) syn-
thetic tractability, and 5) a lack of known macromolecular
targets. Importantly, other publically available target predic-
[4]
chemical biology. Herein, we disclose the targets of the
natural products goitrin, isomacroin, and graveolinine, which
we discovered by using a ligand-based computational method
for predicting the targets of (de-orphaning) NPs in the
absence of apparent chemotype similarity to known macro-
molecular target effectors.
Chemocentric approaches for target prediction have
[
5]
already proven their applicability to drug repurposing,
[
6]
identifying drug liabilities, and driving innovative multi-
[
*] Dr. T. Rodrigues, D. Reker, J. Kunze, Dr. P. Schneider,
Prof. Dr. G. Schneider
Department of Chemistry and Applied Biosciences
Swiss Federal Institute of Technology (ETH)
Vladimir-Prelog-Weg 4, 8093 Zurich (Switzerland)
E-mail: gisbert.schneider@pharma.ethz.ch
[
**] This research was financially supported by ETH Zurich and a grant
from the OPO Foundation, Zurich, Switzerland. P. S. and G. S. are
the founders of inSili.com GmbH, Zurich.
[5e]
[13]
[14]
tion tools, for example, SEA, PASS, and SuperPred,
failed to confidently predict either these or any other targets
for the selected NPs (see the Supporting Information), thus
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201504241.
1
0516
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10516 –10520

Angewandte
Chemie
was obtained in one step via Friedländer chemistry from the
required starting materials. Activation of 4-methoxyquinoline
to its N-oxide counterpart 3 with either mCPBA or MeReO3/
H O oxidation systems enabled subsequent regioselective
2
2
direct Pd-catalyzed CÀH arylation to intermediate 6, as
[15]
reported previously.
Catalytic hydrogen transfer with 6
gave access to graveolinine, 10.
Racemic goitrin had been reported as a moderate inhib-
[
16]
itor of dopamine b-hydroxylase. According to our software,
dl-goitrin was confidently predicted as a pregnane X receptor
(PXR) ligand (Table 1) when predicting targets for the
tautomer reported in the DNP. Interestingly, target inference
was achieved from the PXR reference structure dl-sulphor-
aphane, which presents no substructure similarity to goitrin
according to the Tanimoto similarity coefficient (T = 0.02). In
c
a functional assay, goitrin presented modest antagonism
against PXR (ca. 20% effect at 150 mm, Figure 2A). Sulphor-
aphane had been previously reported as a potent PXR
antagonist with a half maximal inhibitory concentration (IC )
5
0
[17]
of approximately 12 mm, but in our assay it exhibited only
a similar functional effect to goitrin. Furthermore, from
a constrained acetylcholine analogue, we also predicted and
Figure 1. A) Selected natural products and retrosynthetic analysis.
B) Synthetic pathways. Reagents and conditions: i) KOH, EtOH, reflux;
ii) mCPBA, CH Cl , RT, or MeReO , H O 50%, CH Cl , 08C !RT;
2
2
3
2
2
2
2
t
iii) Bromoaryl, K CO , P Bu ·HBF , Pd(OAc) , PhMe, reflux; iv) H-
2
3
3
4
2
CO NH , Pd/C 10%, MeOH, reflux.
2
4
Figure 2. A) Dot plot with median (bar) for the functional cell-based
PXR assay. dl-goitrin and dl-sulphoraphane were assayed for antago-
nist effect at 150 mm; N=2 (control: clotrimazole, IC50 =7.5 mm).
B) Dot plot with median (bar) for radioligand displacement and
highlighting the unprecedented domain of applicability of our
SOM-based algorithm to extend the NP target space.
While racemic goitrin was purchased for screening, we
devised concise synthetic approaches for the alkaloids
isomacroin and graveolinine (Figure 1A,B). Isomacroin (1)
functional cell-based muscarinic M receptor screening of DL-goitrin at
1
2
50 mm; N=2 (controls: radioligand displacement: pirenzepine,
IC50 =30 nm; agonist: acetylcholine EC50 =1.8 nm; antagonist: pirenze-
pine, IC50 =44 nm).
Table 1: Target predictions for fragment-like natural products and nearest neighbors from the drug reference data.
[a]
Natural product
Structure
Target prediction (p value%)
Nearest neighbor structure
Tc
Pregnane X receptor (0.25)
0.02
0.04
DL-Goitrin
Muscarinic M receptor (0.36)
1
Platelet-derived growth factor receptor (0.25)
0.16
0.13
Isomacroin
Adenosine A receptor (0.34)
3
Cyclooxygenase-2 (0.72)
0.08
0.11
Graveolinine
Serotonin 5-HT2B receptor (0.96)
[
a] Tanimoto similarity calculated from Morgan substructure fingerprints (radius=2, nbits=2048).
Angew. Chem. Int. Ed. 2015, 54, 10516 –10520
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10517

.
Angewandte
Communications
experimentally confirmed weak muscarinic M receptor bind-
ing and engagement by goitrin at 250 mm (Figure 2B).
Although the goitrin-inspired design of small molecules as
directed interactions of the imidazole moiety in isomacroin
with the kinase hinge-binding motif of both PDGFR isoforms.
Interestingly, isomacroin is a substructure of the recently
reported single-digit nanomolar and micromolar inhibitors
1
PXR and M modulators may require substantial medicinal
1
[22]
[23]
chemistry effort, the results highlight the biophoric nature of
this compound and the capabilities of the target-prediction
tool in detecting subtle structural features of fragment-like
NPs that are required for specific target recognition and
modulation.
for PDGFRb
and IKKb,
respectively, thus further
supporting the prevalidation of NP-derived molecular archi-
tectures, the suitability of NP fragments as leads for develop-
ment and, in this specific case, the molecular binding
hypothesis probed with 2.
[18]
Isomacroin is of great pharmacological interest.
In
Finally, we selected graveolinine (10) from Ruta grave-
[
24]
[25]
accordance with the top-confidence target predictions, we
screened it in a functional cell-based assay against the
olens, which possesses antiangiogenic and antitubercular
activities. Our software confidently predicted graveolinine to
bind to the serotonin 5-HT2B receptor and cyclooxygenase-2
(COX-2), as inferred from the structurally unrelated small
adenosine A receptor with a confounding outcome at
3
1
50 mm (18% agonism and 13% antagonism). Despite the
[26]
[27]
structural dissimilarity between isomacroin and its reference
molecules pyrroloindole and diphenylthiophene, respec-
tively. Efficient binding and potent serotonin 5-HT2B receptor
[19]
compound (T = 0.16), we successfully revealed isomacroin
c
to be a potent and ligand-efficient inhibitor (IC = 25 mm Æ
antagonism were experimentally confirmed (binding: K =
5
0
i
0
.08 log units; ligand efficiency LE = 0.40; Table 1 and
7 mm Æ 0.14 log units; LE = 0.34; antagonism: IC = 12 mm Æ
50
Figure 3) of platelet-derived growth factor receptor alpha
PDGFRa). Similarly to the reference quinoline, no selectiv-
0.07 log units; Figure 4A) together with an absence of
colloidal aggregation at relevant bioactive concentrations.
Analysis of known 5-HT2B ligands in ChEMBL revealed the
equipotent alkaloid berbevine as its nearest neighbor
(
(
CHEMBL12089; K = 2.2 mm; T = 0.31, Morgan finger-
i
c
prints), thus endorsing the feasibility of using target predic-
tion tools to identify potentially attrition-prone scaffolds,
since the 5-HT2B receptor may be considered as an undesired
[7b]
off-target for drug development.
Graveolinine was also confirmed to engage COX-2 at
1
50 mm (79% effect inhibition; Figure 4B). Extracts of Ruta
graveolens had previously shown antiplatelet aggregation
Figure 3. Concentration–response curve for isomacroin against
PDGFRa. IC50 =24.5 mmÆ0.08 log units (ligand efficiency: LE=0.40);
N=2 (control: staurosporine, IC50 =0.34 nm).
ity between the a and b isoforms was observed at 150 mm
(
35% inhibition of PDGFRb). To the best of our knowledge,
PDGFR kinase constitutes the first disclosed target of
isomacroin. Kinase panel screening revealed more than
5
0% inhibition by isomacroin (25 mm) against only three
(
MNK2 > MAP4K4 > SIK) of 45 kinases assayed, which is in
accordance with the predictions (see the Supporting Infor-
mation). This outcome suggests the isomacroin scaffold as
a candidate for the design of selective kinase inhibitors. We
ruled out artifact and false-positive readout measurements by
confirming the absence of colloidal aggregates at concen-
trations as high as 200 mm (see the Supporting Information).
[20]
Notably, the nearest neighbor in ChEMBL according to
Morgan fingerprints, AG-1295 (CHEMBL7724), presents
identical activity to isomacroin against PDGFRa (42%
[
21]
inhibition at 10 mm) yet low substructure similarity (T =
c
0
.27). These results support the feasibility of discovering
Figure 4. A) Concentration-dependent binding and antagonism by
innovative NP hits for the development of chemical probes
and lead structures without significant loss of target modu-
lation potential compared to structurally unmatched isofunc-
tional molecules.
We further probed the molecular basis of PDGFR kinase
inhibition by synthesizing analogue 2. From the sharply
decreased inhibitory potency of 2 at 150 mm (3% and 10%
inhibition of PDGFRa and b, respectively), our data suggest
graveolinine against the serotonin 5-HT2B receptor (binding:
K
=7 mmÆ0.14 log units; Antagonism: IC50 =12 mmÆ0.07 log units;
i
n=2). B) Dot plot with median (bar) for the inhibition of COX-2 by
graveolinine (10) and its analogues 11–13 at 150 mm; N=2. C) Flexible
alignment superimposition of graveolinine (turquoise) onto its COX-2
[27]
COBRA reference (red); Tanimoto similarity=0.08. D) Docking pose
of 10 in the active site of COX-2. Predicted key interactions are
indicated with dashed lines. Image produced with PyMOL (Schrç-
dinger LLC).
1
0518 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10516 –10520

Angewandte
Chemie
[28]
activity induced by 100 mm arachidonic acid,
which may
[
[
1] a) A. L. Harvey, R. Edrada-Ebel, R. J. Quinn, Nat. Rev. Drug
now be explained by COX-2 inhibition by graveolinine. The
result obtained pinpoints potential competence of the SOM
algorithm in target deconvolution from observations of
phenotype changes induced by complex compound matrices.
In spite of negligible structural similarity to the reference
compound, flexible ligand alignment (MOE v2014.9) with
perfectly matched pharmacophore features corroborates the
retrieval of COX-2 as a target of graveolinine (Figure 4B).
The nearest neighbor of graveolinine in CHEMBL COX-2
Discovery 2015, 14, 111 – 129; b) A. Bauer, M. Bronstrup, Nat.
Prod. Rep. 2014, 31, 35 – 60.
2] a) M. L. Lee, G. Schneider, J. Comb. Chem. 2001, 3, 284 – 289;
b) L. Du, A. J. Robles, J. B. King, D. R. Powell, A. N. Miller, S. L.
Mooberry, R. H. Cichewicz, Angew. Chem. Int. Ed. 2014, 53,
804 – 809; Angew. Chem. 2014, 126, 823 – 828; c) H. van Hattum,
H. Waldmann, J. Am. Chem. Soc. 2014, 136, 11853 – 11859; d) V.
Eschenbrenner-Lux, P. Kuchler, S. Ziegler, K. Kumar, H.
Waldmann, Angew. Chem. Int. Ed. 2014, 53, 2134 – 2137;
Angew. Chem. 2014, 126, 2166 – 2169; e) R. Narayan, J. O.
Bauer, C. Strohmann, A. P. Antonchick, H. Waldmann, Angew.
Chem. Int. Ed. 2013, 52, 12892 – 12896; Angew. Chem. 2013, 125,
[29]
inhibitor space (CHEMBL2023860)
also presents low
structural similarity (T = 0.42). It is important to point out
c
that graveolinine does not possess structural features of
potential assay interference compounds, unlike its
1
3130 – 13134; f) M. C. McLeod, G. Singh, J. N. Plampin 3rd, D.
Rane, J. L. Wang, V. W. Day, J. Aube, Nat. Chem. 2014, 6, 133 –
140; g) T. Miyao, D. Reker, P. Schneider, K. Funatsu, G.
Schneider, Planta Med. 2015, 81, 429 – 435.
[30]
CHEMBL2023860 nearest neighbor, aminothiazole.
We then set out to gain insight into the molecular basis of
COX-2 recognition by graveolinine. Molecular docking into
the active site of COX-2 was carried out with GOLD/
[
3] a) M. A. Koch, A. Schuffenhauer, M. Scheck, S. Wetzel, M.
Casaulta, A. Odermatt, P. Ertl, H. Waldmann, Proc. Natl. Acad.
Sci. USA 2005, 102, 17272 – 17277; b) J. Larsson, J. Gottfries, S.
Muresan, A. Backlund, J. Nat. Prod. 2007, 70, 789 – 794; c) S.
Wetzel, K. Klein, S. Renner, D. Rauh, T. I. Oprea, P. Mutzel, H.
Waldmann, Nat. Chem. Biol. 2009, 5, 581 – 583; d) S. Renner,
W. A. van Otterlo, M. Dominguez Seoane, S. Mocklinghoff, B.
Hofmann, S. Wetzel, A. Schuffenhauer, P. Ertl, T. I. Oprea, D.
Steinhilber, L. Brunsveld, D. Rauh, H. Waldmann, Nat. Chem.
Biol. 2009, 5, 585 – 592; e) B. Over, S. Wetzel, C. Grutter, Y.
Nakai, S. Renner, D. Rauh, H. Waldmann, Nat. Chem. 2013, 5,
[
31]
[32]
GoldScore (v5.2.2) (PDB 4PH9 ) to afford a potential
explanation for the observed functional effects. The binding
120
model suggests key directed interactions with Arg and
3
53
Leu , as well as hydrophobic contacts stemming from the
methoxy group (Figure 4D). Hence, we probed the mode of
binding of graveolinine by synthesizing analogues 11–13. In
accordance with the binding model, compounds 11–13
showed reduced potency at 150 mm compared to graveolinine
2
1 – 28.
(
Figure 4B), thereby corroborating the proposed basis for
[
4] a) D. Reker, A. M. Perna, T. Rodrigues, P. Schneider, M.
Reutlinger, B. Monch, A. Koeberle, C. Lamers, M. Gabler, H.
Steinmetz, R. Muller, M. Schubert-Zsilavecz, O. Werz, G.
Schneider, Nat. Chem. 2014, 6, 1072 – 1078; b) J. M. Rollinger,
D. Schuster, B. Danzl, S. Schwaiger, P. Markt, M. Schmidtke, J.
Gertsch, S. Raduner, G. Wolber, T. Langer, H. Stuppner, Planta
Med. 2009, 75, 195 – 204; c) J. M. Rollinger, Phytochem. Lett.
molecular recognition. Withdrawal of the ether hydrogen-
1
20
bond acceptor that potentially interacts with Arg (com-
pound 11) resulted in marginal effects on activity. In fact,
[33]
aromatic ethers are weak hydrogen-bond acceptors, which
fully supports the obtained biochemical data. Truncation of
the OMe group (compound 12) revealed that this group
significantly contributes to binding efficiency. Interaction
2
009, 2, 53 – 58.
[
5] a) D. Reker, T. Rodrigues, P. Schneider, G. Schneider, Proc. Natl.
Acad. Sci. USA 2014, 111, 4067 – 4072; b) T. Rodrigues, Y.-C.
Lin, M. Hartenfeller, S. Renner, Y. F. Lim, G. Schneider, Chem.
Commun. 2015, 51, 7478 – 7481; c) A. A. Antolín, X. Jalencas, J.
YØlamos, J. Mestres, ACS Chem. Biol. 2012, 7, 1962 – 1967;
d) M. J. Keiser, V. Setola, J. J. Irwin, C. Laggner, A. I. Abbas, S. J.
Hufeisen, N. H. Jensen, M. B. Kuijer, R. C. Matos, T. B. Tran, R.
Whaley, R. A. Glennon, J. Hert, K. L. Thomas, D. D. Edwards,
B. K. Shoichet, B. L. Roth, Nature 2009, 462, 175 – 181; e) M. J.
Keiser, B. L. Roth, B. N. Armbruster, P. Ernsberger, J. J. Irwin,
B. K. Shoichet, Nat. Biotechnol. 2007, 25, 197 – 206.
3
52
with Leu may also be key for enzyme inhibition since
disruption of the suggested CH ···p interaction between
2
3
53
Leu and the quinolyl moiety resulted in a sharp loss of
potency. Altogether, this small focused analogue library
affords a rationale for future derivatization and the design
of graveolinine-inspired COX-2 inhibitors.
With the biological prevalidation of NP architecture in
mind, the results of our present study suggest an expeditious
solution for rapid target identification for NPs. Together with
fragment-based drug design, such platforms could be used for
the efficient and economical design of NP-derived chemical
tools and drug leads, as well as for the analysis of phenotypic
screening hits. The results obtained support the use of target
prediction tools as an enabling technology for the NP-inspired
design of new chemical entities in chemical biology and
molecular medicine.
[6] a) A. A. Antolín, J. Mestres, ACS Chem. Biol. 2015, 10, 395 –
400; b) E. Lounkine, M. J. Keiser, S. Whitebread, D. Mikhailov,
J. Hamon, J. L. Jenkins, P. Lavan, E. Weber, A. K. Doak, S. Cote,
B. K. Shoichet, L. Urban, Nature 2012, 486, 361 – 367.
[
7] a) M. Reutlinger, T. Rodrigues, P. Schneider, G. Schneider,
Angew. Chem. Int. Ed. 2014, 53, 4244 – 4248; Angew. Chem.
2
014, 126, 4330 – 4334; b) T. Rodrigues, N. Hauser, D. Reker, M.
Reutlinger, T. Wunderlin, J. Hamon, G. Koch, G. Schneider,
Angew. Chem. Int. Ed. 2015, 54, 1551 – 1555; Angew. Chem.
2015, 127, 1571 – 1575; c) J. Besnard, G. F. Ruda, V. Setola, K.
Abecassis, R. M. Rodriguiz, X. P. Huang, S. Norval, M. F.
Sassano, A. I. Shin, L. A. Webster, F. R. Simeons, L. Stojanovski,
A. Prat, N. G. Seidah, D. B. Constam, G. R. Bickerton, K. D.
Read, W. C. Wetsel, I. H. Gilbert, B. L. Roth, A. L. Hopkins,
Nature 2012, 492, 215 – 220.
Keywords: chemical biology · computer-assisted drug design ·
drug discovery · medicinal chemistry · target identification
How to cite: Angew. Chem. Int. Ed. 2015, 54, 10516–10520
Angew. Chem. 2015, 127, 10662–10666
[
[
8] C. W. Murray, D. C. Rees, Nat. Chem. 2009, 1, 187 – 192.
9] M. Congreve, R. Carr, C. Murray, H. Jhoti, Drug Discovery
Today 2003, 8, 876 – 877.
Angew. Chem. Int. Ed. 2015, 54, 10516 –10520
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10519

.
Angewandte
Communications
[
[
10] M. Reutlinger, C. P. Koch, D. Reker, N. Todoroff, P. Schneider, T.
Rodrigues, G. Schneider, Mol. Inf. 2013, 32, 133 – 138.
11] a) P. Schneider, Y. Tanrikulu, G. Schneider, Curr. Med. Chem.
C. Parast, Y. Shin, E. B. Sjogren, M. L. Smith, F. X. Talamas, G.
Tonn, K. M. Walker, N. P. Walker, H. Wesche, C. Whitehead, M.
Wright, M. F. Browner, Bioorg. Med. Chem. Lett. 2011, 21, 417 –
422.
2
009, 16, 258 – 266; b) T. Kohonen, Biol. Cybern. 1982, 43, 59 –
6
9.
[24] Z. Y. An, Y. Y. Yan, D. Peng, T. M. Ou, J. H. Tan, S. L. Huang,
L. K. An, L. Q. Gu, Z. S. Huang, Eur. J. Med. Chem. 2010, 45,
3895 – 3903.
[
[
12] P. Schneider, G. Schneider, QSAR Comb. Sci. 2003, 22, 713 – 718.
13] A. Lagunin, D. Filimonov, V. Poroikov, Curr. Pharm. Des. 2010,
1
6, 1703 – 1717.
[25] N. Kishore, B. B. Mishra, V. Tripathi, V. K. Tiwari, Fitoterapia
2009, 80, 149 – 163.
[
[
[
[
14] M. Dunkel, S. Gunther, J. Ahmed, B. Wittig, R. Preissner,
Nucleic Acids Res. 2008, 36, W55 – 59.
15] L. C. Campeau, S. Rousseaux, K. Fagnou, J. Am. Chem. Soc.
[26] R. Hajjo, C. M. Grulke, A. Golbraikh, V. Setola, X. P. Huang,
B. L. Roth, A. Tropsha, J. Med. Chem. 2010, 53, 7573 – 7586.
[27] Y. Leblanc, J. Gauthier, D. Ethier, J. Guay, J. Mancini, D.
Riendeau, P. Tagari, P. Vickers, E. Wong, P. Prasit, Bioorg. Med.
Chem. Lett. 1995, 5, 2123 – 2128.
2
005, 127, 18020 – 18021.
16] N. Zenker, L. S. Hubbard, J. Wright, J. Nat. Prod. 1988, 51, 862 –
65.
8
17] C. Zhou, E. J. Poulton, F. Grun, T. K. Bammler, B. Blumberg,
[28] T.-S. Wu, L.-S. Shi, J.-J. Wang, S.-C. Iou, H.-C. Chang, Y.-P. Chen,
Y.-H. Kuo, Y.-L. Chang, C.-M. Teng, J. Chin. Chem. Soc. 2003,
50, 171 – 178.
K. E. Thummel, D. L. Eaton, Mol. Pharmacol. 2007, 71, 220 –
2
29.
[
[
18] A. Jordaan, R. R. Arndt, J. Heterocycl. Chem. 1968, 5, 723 – 725.
19] M. P. Maguire, K. R. Sheets, K. McVety, A. P. Spada, A.
Zilberstein, J. Med. Chem. 1994, 37, 2129 – 2137.
[29] B. Smith, H. H. Chang, F. Medda, V. Gokhale, J. Dietrich, A.
Davis, E. J. Meuillet, C. Hulme, Bioorg. Med. Chem. Lett. 2012,
22, 3567 – 3570.
[
20] A. P. Bento, A. Gaulton, A. Hersey, L. J. Bellis, J. Chambers, M.
Davies, F. A. Kruger, Y. Light, L. Mak, S. McGlinchey, M.
Nowotka, G. Papadatos, R. Santos, J. P. Overington, Nucleic
Acids Res. 2014, 42, D1083 – 1090.
[30] a) S. M. Devine, M. D. Mulcair, C. O. Debono, E. W. Leung,
J. W. Nissink, S. S. Lim, I. R. Chandrashekaran, M. Vazirani, B.
Mohanty, J. S. Simpson, J. B. Baell, P. J. Scammells, R. S. Norton,
M. J. Scanlon, J. Med. Chem. 2015, 58, 1205 – 1214; b) J. B. Baell,
G. A. Holloway, J. Med. Chem. 2010, 53, 2719 – 2740.
[31] G. Jones, P. Willett, R. C. Glen, J. Mol. Biol. 1995, 245, 43 – 53.
[32] B. J. Orlando, M. J. Lucido, M. G. Malkowski, J. Struct. Biol.
2015, 189, 62 – 66.
[
21] Y. Gao, S. P. Davies, M. Augustin, A. Woodward, U. A. Patel, R.
Kovelman, K. J. Harvey, Biochem. J. 2013, 451, 313 – 328.
22] E. J. Hicken, F. P. Marmsater, M. C. Munson, S. T. Schlachter,
J. E. Robinson, S. Allen, L. E. Burgess, R. K. DeLisle, J. P. Rizzi,
G. T. Topalov, Q. Zhao, J. M. Hicks, N. C. Kallan, E. Tarlton, A.
Allen, M. Callejo, A. Cox, S. Rana, N. Klopfenstein, R.
Woessner, J. P. Lyssikatos, ACS Med. Chem. Lett. 2014, 5, 78 – 83.
23] T. D. Cushing, V. Baichwal, K. Berry, R. Billedeau, V. Bordunov,
C. Broka, M. Cardozo, P. Cheng, D. Clark, S. Dalrymple, M.
DeGraffenreid, A. Gill, X. Hao, R. C. Hawley, X. He, J. C. Jaen,
S. S. Labadie, M. Labelle, C. Lehel, P. P. Lu, J. McIntosh, S. Miao,
[
[33] C. Bissantz, B. Kuhn, M. Stahl, J. Med. Chem. 2010, 53, 5061 –
5084.
[
Received: May 9, 2015
Revised: June 4, 2015
Published online: July 17, 2015
1
0520 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10516 –10520