The first de novo-designed antagonists of the human NK2 receptor

Article abstract of DOI:10.1021/jm050533o

The de novo molecular design program SPROUT has been used in conjunction with a molecular model to produce a molecular template for a new class of NK₂ receptor antagonist. An efficient, stereocontrolled synthesis of a small series of molecules,

Full text of DOI:10.1021/jm050533o

J. Med. Chem. 2005, 48, 5655-5658
5655
The First de Novo-Designed Antagonists
of the Human NK2 Receptor
Mohammed A. Ali,† Nirmala Bhogal,‡
John B. C. Findlay,‡ and Colin W. G. Fishwick*,†
School of Chemistry, and Department of Biochemistry and
Molecular Biology, University of Leeds, Leeds LS2 9JT, UK
Figure 1. Examples of known NK2 antagonists.
Received June 7, 2005
mutants has suggested that these two classes of ligands
bind to nonidentical receptor sites which overlap par-
tially near the extra-cellular portions of helices VI and
Abstract: The de novo molecular design program SPROUT
has been used in conjunction with a molecular model to
produce a molecular template for a new class of NK2 receptor
antagonist. An efficient, stereocontrolled synthesis of a small
series of molecules, designed to test the validity of this
template, was developed. Competition assays using recombi-
nant human NK2 receptor support the structural requirements
of this new designed molecular template.
14-15
VII.
In the absence of X-ray crystal structural data for
these receptors, a 3D molecular model has been devel-
oped16 which utilizes the projection structure of the
light-activated GPCR rhodopsin as a template.17
The resultant model accommodates data derived from
sequence analysis algorithms for the putative trans-
membrane segments and indicates the relative orienta-
tion of amino acid residues within each R helix. We have
previously applied a computational automated docking
procedure to this model in order to identify the non-
peptidic antagonist binding region.18 This was predicted
to be a broadly cylindrical region spanning the space
enclosed by helices III, VI, and VII and bounded by
residues N86 (helix II), Q109 and A116 (helix III), H198
(helix V), Y266 (helix VI), and W294 (helix VII) (Figure
2). Evidence supporting the validity of this proposed
binding region was obtained from a study of the binding
affinities of a series of ligands based on 2 which were
designed to test the predicted mode of binding of 2 to
the receptor.
The G protein-coupled receptor (GPCR) superfamily
constitute perhaps the most ubiquitous signal trans-
duction system in eukaryotic species. This family are
one of the largest and most diverse groups of proteins
in humans and are encoded by up to 3% of genes within
the human genome. These receptors mediate vision,
smell, taste, neurotransmission, hormonal and immune
responses, and even viral infection.1 It is therefore not
surprising that the central role of GPCRs in cellular
signaling renders them primary targets for drug therapy.
Indeed, recent estimates suggest that up to 60% of
modern pharmacoepia is targeted to GPCRs.2 These
integral membrane proteins all possess a heptahelical
intramembranous bundle interconnected by extracellu-
lar and intracellular hydrophilic segments.3-5
The neurokinin (NK) receptors are an important class
of GPCR that are located in the central nervous system,
peripheral neurones, and smooth muscle tissue. There
are at least three known NK receptors, NK1, NK2 and
NK3, although various subtypes and species differences
have been discovered.3 The naturally occurring neu-
ropeptides substance P, NKA, and NKB are potent
agonists for each of these receptors, respectively. The
interaction of these peptides with their respective
receptors is implicated in a range of biological responses
including pain transmission, inflammation, and Par-
kinson’s and Alzheimer’s diseases. There is therefore
tremendous potential for the development of highly
specific, bioavailable small molecule antagonists.
To date, in addition to a number of peptidic antago-
nists,6,7 there are currently two main classes of non-
peptidic NK2 receptor antagonists (Figure 1); the ben-
zamide piperidine antagonists, e.g., 1,8-12 and those
derived from indolium piperidines, e.g., 213 (Figure 1).
Although much elegant work has been done to refine
the structure-activity relationships within these classes,
essentially all of the known antagonists have resulted
from compound library screening methods. Pharmaco-
logical characterization of site-specific NK2 receptor
The structure-based design of small-molecule enzyme
inhibitors is a powerful tool in drug discovery, particu-
larly when there is an X-ray crystal structure of the
target enzyme available.
Recent reports involving the application of de novo
design indicate that this approach is a powerful alterna-
tive to these classical drug discovery approaches.19-23
Here, by using the structural features present within
the enzyme only, new inhibitor designs are built-up
sequentially according to the requirements of the tar-
geted binding site. Therefore, in principle, de novo
design should be a far more useful technique than
virtual high-throughput screening, because a good de
novo design program will examine structure space
larger by many orders of magnitude than that of most
virtual libraries currently used for this purpose.
To design a completely new class of NK2 antagonist,
we decided to apply a de novo design approach to this
putative antagonist binding cavity. We were aware that
such an objective, based solely upon a molecular model
of a biological target, as opposed to an X-ray or NMR-
derived structure, was ambitious. Nevertheless, such a
structure-based approach offers tremendous potential
in terms of the ability to tailor the design to increase
antagonist potency and selectivity for a particular
receptor.
* To whom correspondence should be addressed. Tel. +44(0)113-
3436510; Fax. +44(0)113-3436565; Email: colinf@chem.leeds.ac.uk.
† School of Chemistry.
Here we report the application of the de novo molec-
‡ Department of Biochemistry and Molecular Biology.
ular design program SPROUT24 to the design of novel
10.1021/jm050533o CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/11/2005

5656 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Letters
due to their close proximity to the hydrophobic regions.
The resulting designed antagonist template consisted
of a five-membered cyclic core possessing two chains,
1,3-disposed, and terminating in hydrophobic groups
designed to interact with the hydrophobic regions within
the cavity. Additionally, the design called for the pres-
ence of two specifically located heteroatoms. One of
these forms part of one of the hydrophobic chains and
accepts an H-bond from Q109 helix III. The other is
located as part of a small chain on the central frame-
work. Both of the hydrophobic chains and this smaller
chain are required to have an all-cis relationship (Figure
3).
We have previously reported a simple and convenient
method for the synthesis of 1,3-cis disubstituted tet-
rahydrofuranyl rings based upon the oxidative ring-
opening of the adducts derived from cycloadditions of
3-oxidopyrilium betaines.26 We reasoned that molecules
possessing the structure required by the designed
template could be readily accessed using this chemistry,
and a short synthesis of a range of such systems 6
designed to test the validity of the antagonist template
in Figure 1 is shown below (Scheme 1).
Ketone protection of bicycle 3 (obtained along with
the C-3 diastereoisomeric ester as a separable 1:1 mix-
ture; total yield 76%, from cycloaddition of 3-oxidopy-
rilium betaine with methyl acrylate using the literature
procedure27) followed by ester reduction and protection
gave ether 4 in good yield. Ozonolytic cleavage and
reduction then gave diol 5 which was directly alkylated
to yield ethers 6. Use of the more bulky 2-(bromometh-
yl)naphthalene as alkylating agent resulted in predomi-
nately monoalkylation on the chain attached to C-5 of
the THF ring. This selectivity of alkylation was used to
prepare ethers 6 which possessed differing aromatic
substituents. An analogous reaction sequence applied
to the C-3 epimeric ester of compound 3 was also used
to give access to the C-3 epimeric alcohol corresponding
to 6. We were surprised to find that all attempts at
hydrolysis of ketals 6 to reveal the corresponding
ketones were unsuccessful, mild acid hydrolysis failing
completely and more vigorous acid hydrolysis or the use
of Lewis acids giving complex reaction mixtures.
To ascertain the likely effect of incorporation of the
ketal unit upon receptor binding, the computer model
containing the designed template was modified so as to
incorporate the ketal unit. Following energy minimiza-
tion within the receptor model, comparison of this
structure with the original template revealed that the
ketal moiety is located in a space near to helix VII
between M297 and S299. Although no additional bond-
ing interactions are present, this modification has very
little effect on the positioning of the original template
(Figure 4), and we were satisfied that structures 6 would
be a good representation of the SPROUT-designed
antagonist template.
Figure 2. Model of transmembrane region of the NK2 receptor
showing putative antagonist binding region containing 2. (Top)
View looking down into receptor from extracellular side
(antagonist shaded pink). (Bottom) Side-view with ligand
represented in space-fill and extracellular section of receptor
facing uppermost.
antagonists of the human tachykinin NK2 receptor. We
also describe the development of efficient and stereo-
selective syntheses of these antagonists as well as their
biological evaluation.
SPROUT24 is a powerful suite of software modules for
de novo structure-based molecular design. SPROUT has
modules for (i) characterizing regions within a protein,
(ii) detecting ‘hotspots’ where ligand atoms would form
strong interactions with the protein, (iii) docking mo-
lecular fragments to these sites, (iv) joining these
fragments with a molecular backbone to give a complete
ligand, and (v) ranking the set of predicted ligands by
criteria including predicted binding energy, structural
complexity, and synthetic accessibility.
To utilize a de novo approach for the design of new
NK2 antagonists, we applied SPROUT to the putative
antagonist binding region within our model25 of the NK2
receptor.
To design structurally simple antagonists which were
predicted to show reasonably good affinity for the
enzyme, we wished to utilize only a small number of
antagonist-receptor contacts in our SPROUT design
strategy. Our earlier work had revealed the presence
of two regions of relatively high hydrophobicity that we
wished to utilize in antagonist design. Additionally,
SPROUT analysis revealed the presence of a number
of potential H-bonding sites within the putative binding
cavity. From these, a side-chain NH from Q109 helix
III (termed a ‘donor’ site) and a backbone carbonyl from
T300 helix VII (termed an ‘acceptor’ site) were chosen
Biological evaluation28 of the designed antagonists 6
was performed using competition assays with [3H]-
SR48968 1 and recombinant human NK2 (Table 1).
As can be seen from Table 1 entry 1, the compound
corresponding to the SPROUT-designed antagonist
template shows the greatest affinity for the NK2 recep-
tor. Changing the stereochemistry at the C-3 atom in
structure 6 results in greater than a 500-fold decrease

Letters
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5657
Scheme 1. Synthesis of Designed Antagonists 6a
a Reagents and conditions: (a) ethylene glycol, TsOH, 61%; (b)
LiAlH4, Et2O, 72%; (c) t-BuMe2SiCl, DMF, 71%; (d) O3, MeOH;
(e) NaBH4, MeOH; (f) NaH, THF; (g) ArCH2Br, THF, 53-65% over
steps d-g; (h) nBu4NF, THF, 87-98%.
Figure 4. Overlay of SPROUT-designed template (darker
shading) with ketal-modified system (lighter shading), both
located within the NK2 receptor model (not shown).
Table 1. Structures and Activities of Designed Antagonists
Figure 3. SPROUT-generated NK2 antagonist template. (Top)
The hydrophobic cavities (yellow) and the H-bonding sites (blue
and red) utilized in the design. (Middle) Schematic showing
contacting residues (X ) N or O). (Bottom) Overlay of designed
template and docked ligand 2 (shaded in pink).
in activity (entry 2), in keeping with the requirement
for an ‘all-cis’ stereochemical relationship.
Interestingly, increasing the bulk of the hydrophobic
substituent attached to C-5 of the ring from phenyl to
naphthyl (entry 4) reduces the affinity by greater than
200-fold whereas changing both hydrophobic substitu-
ents to naphthyl leads to more than a 300-fold decrease
in affinity (entry 3). These observations appear to
indicate the size-constraints within the hydrophobic
cavities.
a Obtained from the C-3 diastereoisomer of 3 using the synthetic
sequence shown in Scheme 1. b Obtained as racemic mixtures.
chosen to perform biological evaluation using readily
available enantiomeric mixtures containing this de-
signed template along with the enantiomer. Therefore,
although the measured binding affinities of these mix-
tures to NK2 are very encouraging, it is likely that the
enantiomerically pure versions corresponding to the
designed molecular template are even more potent in
the presence of NK2.
Clearly, the designed template is for a single enan-
tiomer of absolute stereochemistry shown in Figure 3.
In these proof-of-principal studies, however, we have

5658 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Letters
(13) Cooper, A. J. W.; Adams, H. S.; Bell, R.; Gore, P. M.; McElroy,
A. B.; Pritchard, J. M.; Smith, P. W.; Ward, P. GR159897 and
related analogs as highly potent, orally active nonpeptide
neurokinin NK2 receptor antagonists. Bioorg. Med. Chem. Lett,
1994, 4, 1951-1956.
(14) Giolitti, A.; Cucchi, P.; Renzetti, A. R.; Rotondaro, L.; Zappitelli,
S.; Maggi, C. A. Molecular determinants of peptide and nonpep-
tide NK-2 receptor antagonists binding sites of the human
tachykinin NK-2 receptor by site-directed mutagenesis. Neu-
ropharmacology 2000, 39, 1422-1429.
(15) Gether, U.; Emonds-Alt, X.; Breliere, J. C.; Fujii, T.; Hagiwara,
D.; Pradier, L.; Garret, C.; Johnsen, T. E.; Schwartz, T. W.
Evidence for a common molecular-mode of action for chemically
distinct nonpeptide antagonists at the neurokinin-1 (substance-
P) receptor. Mol. Pharmacol. 1994, 45, 500-508.
(16) Donnelly, D.; Maudsley, S.; Gent, J. P.; Moser, R. N.; Hurrell,
C. R.; Findlay, J. B. C. Conserved polar residues in the
transmembrane domain of the human tachykinin NK2 recep-
tor: functional roles and structural implications. Biochem. J.
1999, 339, 55-61.
In conclusion, the de novo design program SPROUT
has been used in conjunction with a stereoselective
synthetic route, to produce a simple molecular template
for a new class of NK2 antagonist. We believe that this
methodology is complementary to the use of high-
throughput screening and is particularly attractive
where such screening methodology is not available or
where access to large collections of library compounds
of sufficient molecular diversity is limited. This ap-
proach clearly holds tremendous potential for the iden-
tification of useful leads for the development of new
GPCR-targeted drugs.
Supporting Information Available: Experimental pro-
cedures, including analytical, spectral, and X-ray crystal-
lographic data for the preparation of antagonists 6, and details
of the NK2 receptor binding assays. This material is available
free of charge via the Internet at http://pubs.acs.org.
(17) Unger, U. V.; Hargrave, P. A.; Baldwin, J. M.; Schertler, G. F.
X. Arrangement of rhodopsin transmembrane alpha-helices.
Nature 1997, 389, 203-206.
(18) Ali, M. A.; Bhogal, N.; Fishwick, C. W. G.; Findlay, J. B. C.
Spatial requirements of the antagonist binding site of the NK2
receptor. Bioorg. Med. Chem. Lett. 2001, 11, 819-822.
(19) Branson, K. M.; Smith, B. J. The role of virtual screening in
computer aided structure-based drug design. Aust. J. Chem.
2004, 57, 1029-1037.
(20) Schneider, G.; Clement-Chomienne, O.; Hilfiger, L.; Schneider,
P.; S.; Kirsch, S.; Bohm, H.-J.; Neidhart, W. Virtual screening
for bioactive molecules by evolutionary de novo design. Angew.
Chem., Int. Ed. 2000, 39, 4130-4133.
(21) Han, Q.; Dominguez, C.; Stouten, P. F. W.; Park, J. M.; Duffy,
D. E.; Galemmo, R. A.; Rossi, K. A.; Alexander, R. S.; Smallwood,
A. M.; Wong, P. C.; Wright, M. M.; Luettgen, J. M.; Knabb, R.
M.; Wexler, R. R. Design, synthesis, and biological evaluation
of potent and selective amidino bicyclic factor Xa inhibitors. J.
Med. Chem. 2000, 43, 4398-4415.
(22) Honma, T.; Haysashi, K.; Aoyama, T.; Hashimoto, N.; Machida,
T.; Fukasawa, K.; Iwama, T.; Ikeura, C.; Ikuta, M.; Suzuki-
Takahashi, I.; Iwasawa, Y.; Hayama, T.; Nishimura, S.; Mor-
ishima, H. Structure-based generation of a new class of potent
Cdk4 inhibitors: new de Novo design strategy and library
design. J. Med. Chem. 2001, 44, 4615-4627.
(23) Grzybowski, B. A.; Ishchenko, A. V.; Kim, C. Y.; Topalov, G.;
Chapman, R.; Christianson, D. W.; Whitesides, G. M.; Shakh-
novich, E. I. Combinatorial computational method gives new
picomolar ligands for a known enzyme. Proc. Natl. Acad. Sci.
U.S.A.. 2002, 99, 1270-1273.
(24) Gillett, V. J.; Newell, W.; Mata, P.; Myatt, G.; Sike, S.; Zsoldos,
Z.; Johnson, A. P. SPROUT-recent developments in the de novo
design of molecules. J. Chem. Inf. Comput. Sci. 1994, 34, 207-
217.
References
(1) Fauci, A. S. Host factors and the pathogenesis of HIV-induced
disease. Nature 1996, 384, 529-534.
(2) Guderman, T.; Neurenberg, B.; Schultz, G. Receptors and
G-proteins as primary components of transmembrane signal
transduction. J. Mol. Med. 1995, 73, 51-63.
(3) Maggi, C. A. The troubled story of tachykinins and neurokinins.
Trends Pharmacol. Sci. 2000, 21 173-175.
(4) Horn, F.; Wear, J.; Beulers, M. W.; Horsch, S.; Baurock, A.; Chen,
W.; Edvarsden, O.; Campagne, F.; Vriend, G. GPCRDB: an
information system for G protein-coupled receptors. Nucleic
Acids Res. 1998, 26, 275-279.
(5) Baldwin, J. M. The probable arrangement of the helices in
G-protein-coupled receptors. EMBO J. 1993, 12, 1693-1703.
(6) Catalioto R. M.; Criscuoli, M.; Cucchi, P.; Giachetti, A.; Giannotti,
D.; Giuliani, S.; Lecci, A.; Lippi, A.; Patacchini, R.; Quartara,
L.; Renzetti, A. R.; Tramontana, M.; Arcamone, F.; Maggi, C. A.
MEN 11420 (Nepadutant), a novel glycosylated bicyclic peptide
tachykinin NK2 receptor antagonist. Br. J. Pharmacol. 1998,
123, 81-91.
(7) Maggi, C. A.; Astolfi, M.; Giuliani, S.; Goso, C.; Manzini, S.;
Meini, S.; Patacchini, R.; Pavone, V.; Pedone, C.; Quartara, L
MEN-10, 627, A novel polycyclic peptide antagonist of tachykinin
NK2 receptors.. J. Pharm. Exp. Ther. 1994, 271, 1489-1500.
(8) Emonds-Alt, X.; Advenier, C.; Croci, T.; Manara, L.; Poncelet,
M.; Proietto, V.; Neliat, G.; Breliere, J. C.; Proietto, V.; Vilain,
P.; Van Broeck, D.; Le Fur, G.; Santucci, V.; Soubre, P.; SR-
48968, A neurokinin-A (NK2) receptor antagonist. Regul. Pept.
1993, 46, 31-36.
(9) Emonds-Alt, X.; Vilain, P.; Cognon, C.; Croci, T.; Proietto, V.;
Van Broeck, D.; Advenier, C.; Naline, E.; Neliat, G.; Le Fur, G.;
Breliere, J. C.; Doucoux, J. P.; Landi, M.; Poncelet, M.; Soubre,
P. Biochemical and pharmacological activities of SR 144190, a
new potent non-peptide tachykinin NK2 receptor antagonist.
Neuropeptides 1997, 31, 449-458.
(10) Kubota, H.; Kakefuda, A. B.; Okamoto, Y.; Fulii, M.; Yamamoto,
O.; Yamagiwa, Y.; Orita, M.; Ikeda, K. Spiro-substituted pip-
eridines as neurokinin receptor antagonists. III. Synthesis of (()-
N-[2-(3,4-dichlorophenyl)-4-(spiro-substituted piperidin-1′-yl)-
butyl]-N-methylbenzamides and evaluation of NK1-NK2 dual
antagonistic activities. Pharmacol. Bull. 1998, 46, 1538-1544.
(11) Jacobs, R. T.; Shenvi, A. B.; Mauger, R. C.; Ulatowski, T. G.;
Aharony, D.; Buckner, C. K. 4-Alkylpiperidines related to
SR48968: Potent antagonists of the neurokinin-2 (NK2) recep-
tor. Bioorg. Med. Chem. Lett. 1998, 8, 1935-1940.
(25) For the present work, the earlier model was refined using the
single-crystal X-ray structure of bovine rhodopsin; see Teller,
D. C.; Okada, T.; Behnke, C. A.; Palczewski, K.; Stenkamp, R.
E. Advances in determination of
a high-resolution three-
dimensional structure of rhodopsin, a model of G-protein-coupled
receptors (GPCRs). Biochemistry 2001, 40, 7761-7772. In
particular, small adjustments in the positioning of helices II,
III, VI, and VI were made in order to better align these with
the corresponding helices in the rhodopsin crystal structure.
(26) Fishwick, C. W. G.; Mitchell, G.; Pang, P. F. W. Ozonolytic
cleavage of 3-oxidopyrilium betaine-derived cycloadducts: Simple
and efficient access to stereodefined tetrahydrofurans. Synlett
2005, 285-286.
(27) Sammes, P. G.; Street, L. J. The preparation and some reactions
of 3-oxidopyrilium. J. Chem. Soc., Perkin Trans. 1 1983, 1261-
1265.
(28) Assays performed as described in reference 12 and in the
Supporting Information.
(12) Kersey, I. D.; Bhogal, N.; Donnelly, D.; Fishwick, C. W. G.;
Findlay, J. B. C.; Ward, P. A non-peptidic photoactivatable
antagonist for mapping the antagonist binding site of the
tachykinin NK2 receptor. Bioorg. Med. Chem. Lett. 1996, 6, 605-
608.
JM050533O