Identification of new potent GPR119 agonists by combining virtual screening and combinatorial chemistry

Article abstract of DOI:10.1021/jm301549a

Virtual screening in a huge collection of virtual combinatorial libraries has led to the identification of two new structural classes of GPR119 agonists with submicromolar in vitro potencies. Herein, we describe the virtual screening process involving feature trees fragment space searches followed by a 3D postprocessing step. The in silico findings were then filtered and prioritized, and finally, combinatorial libraries of target molecules were synthesized. Furthermore the so-called activity-anchor principle is introduced as an element to increase the chance to generate true hits. An activity anchor is a structural element expected to provide key contributions to a certain biological activity. Application of this technique has led to the discovery of two new GPR119-agonist hit series, one of which was further optimized to progress as a novel lead class.

Full text of DOI:10.1021/jm301549a

Article
pubs.acs.org/jmc
Identification of New Potent GPR119 Agonists by Combining Virtual
Screening and Combinatorial Chemistry
Bernd Wellenzohn,† Uta Lessel,† Andreas Beller,‡ Timo Isambert,§ Christoph Hoenke,‡
and Bernd Nosse*,§
†Research Germany/Lead Identification and Optimization Support, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer
Straße 65, 88397 Biberach an der Riss, Germany
‡Research Germany/Medicinal Chemistry/Combinatorial Chemistry, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer
Straße 65, 88397 Biberach an der Riss, Germany
§Research Germany/Medicinal Chemistry, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer Straße 65, 88397 Biberach
an der Riss, Germany
ABSTRACT: Virtual screening in a huge collection of virtual
combinatorial libraries has led to the identification of two new
structural classes of GPR119 agonists with submicromolar in
vitro potencies. Herein, we describe the virtual screening
process involving feature trees fragment space searches
followed by a 3D postprocessing step. The in silico findings
were then filtered and prioritized, and finally, combinatorial
libraries of target molecules were synthesized. Furthermore the
so-called “activity-anchor principle” is introduced as an
element to increase the chance to generate true hits. An
activity anchor is a structural element expected to provide key
contributions to a certain biological activity. Application of this
technique has led to the discovery of two new GPR119-agonist hit series, one of which was further optimized to progress as a
novel lead class.
has already been published14 and is similar to a database
INTRODUCTION
■
reported by Pfizer.15 BICLAIM is based on combinatorial
In the year 2000 at least 171 million people in the world
suffered from type 2 diabetes mellitus (T2DM), and it is
reactions and stored in a feature trees (FTrees) fragment space.
When the virtual screening (VS) campaign for GPR119
expected that the incidence of this disease will almost double
agonists was performed, the virtual chemical space contained
within just one generation to reach 335 million in 2025.1
about 1600 scaffolds and 30 000 reagents encoding about 5 ×
Consequently there is an enormous need for innovative drugs
1011 theoretically chemically accessible molecules. Our
in this field. A characteristic feature of the progression of
T2DM is a gradual loss of pancreatic β-cell function.2 The G-
continuous efforts to expand the BICLAIM space by extracting
protein-coupled receptor GPR119 was recently deorphanized3
in-house and external (data) sources have led to approximately
41 300 scaffolds today. For each library scaffold, possible
and is a focus of current drug development as a new
antidiabetic target.4−7 The receptor is expressed in pancreatic
synthesis routes and lists of reagents compatible with the
β-cells as well as in the gastrointestinal tract,7 and GPR119
agonists have been shown to promote the glucose dependent
corresponding chemistry providing the decoration of the cores
have been collected and stored.
insulin secretion.5 Furthermore it has been reported that
The advantage of concentrating on combinatorially accessible
space is that this chemistry is usually very robust and fast. Thus,
GPR119 agonists stimulate the secretion of glucagon-like
peptide 1 (GLP-1) from intestinal L-cells.5 GPR119 agonists
interesting virtual hits can be easily synthesized and explored by
combinatorial chemistry. We rely on synthetic chemistry which
support a very interesting new drug concept, as they are
expected to regulate the glucose homeostasis by different
mechanisms. Several GPR119 agonists have been published,
is well established preferably in-house or in literature as well as
on readily available building blocks for our combinatorial
chemistry efforts to ensure fast alignment with other hit
identification activities (e.g., HTS or fragment based screening,
and some compounds have already reached the stage of clinical
trials.8−13
FBS). Although the VS techniques enrich active molecules in
In our group, similarity searches in a huge virtual library have
the hit set, very similar molecules have a problem of inaccurate
been performed in addition to high-throughput screening
(HTS) with the aim to identify new GPR119 agonists. Our
virtual library called BICLAIM (Boehringer Ingelheim
Comprehensive Library of Accessible Innovative Molecules)
Received: October 23, 2012
Published: December 4, 2012
© 2012 American Chemical Society
11031
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
Figure 1. Workflow of FTrees fragment space searches at Boehringer Ingelheim (BICLAIM campaign).
predictivity of their biological activity in virtual screening and
there is no inherent solution to avoid false positives. Therefore,
it is not recommended to synthesize and test exclusively the few
best predicted molecule(s) but a larger number of analogues in
a combinatorial library. This enhances the chance to get active
hits and additionally allows for a fast assessment of structure−
activity-relationships (SARs). The SAR information obtained in
the first generation of molecules can be used for the design of
follow-up libraries during an exploration and/or optimization
phase. In our hands the combination of virtual screening and
combinatorial chemistry has been very successful in identifying
new lead structures for several projects. Here we describe the
application of this technique for the discovery of new GPR119
agonists. Two new classes with submicromolar in vitro potency
could be identified.
comparison of the nodes’ topologies. The fragment space
module18 of the FTrees software is making use of the additivity
of the FTrees similarity and thus allows searching a large virtual
(combinatorial) space. This approach is very fast, as only the
fragments representing the potential parts of a molecule from
combinatorial chemistry have to be considered instead of
comparing enumerated molecules.
The resulting hits are then prioritized by a project-specific
postprocessing workflow (steps 3 and 4 in Figure 1). As a
consequence, interesting similar molecules containing all
essential pharmacophoric features can be successfully identified
by FTrees fragment space searches, but the method does not
guarantee an optimal arrangement in space. We postprocess the
FTrees results by 3D virtual screening techniques like docking
or 3D alignments to filter and identify those hits that are also
similar to the query molecules in 3D (see step 3 in Figure 1).
Finally, the summarized results are visually inspected by a team
of medicinal, computational, and combinatorial chemists with
the goal to select the most promising cores (step 5 in Figure 1).
The next step benefits from the fragment space being based on
combinatorial chemistry. Instead of synthesizing only the few
“optimal” FTrees hits, we design and synthesize combinatorial
libraries usually comprising 50−1000 compounds. In many
cases we observed that the decoration that is suggested by in
silico methods to be optimal for the query molecules indeed
resulted in active hits. However, these hits usually need further
optimization, e.g., by improved decoration of the core or
slightly modified scaffolds in order to reach optimal activity.
For the design of focused combinatorial libraries different
options can be used. We apply the software tool LoFT which
has been developed for the design of focused combinatorial
RESULTS AND DISCUSSION
■
Virtual Screening. The outline of the BICLAIM workflow
is illustrated in Figure 1.
The virtual library BICLAIM is stored in a FTrees fragment
space which allows for fast similarity searches against given
queries based on the FTrees descriptors.16−18 The FTrees
software cleaves all rotatable single bonds and represents the
resulting fragments by nodes that are connected in the same
manner like the fragments. Each node is annotated with steric
properties of the corresponding fragments like for instance its
volume and an interaction profile with the H-bond donor and
acceptor capabilities.16 In order to compare two molecules
based on FTrees descriptors, a FTrees score describing the
similarity of the properties of the corresponding nodes is
maximized. This involves a conformation independent
11032
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
Figure 2. Query molecules and their superimposition (1 white, 2 green, 3 blue, 4 yellow).
libraries based on feature trees similarity and physicochemical
properties.19−21 However, for the application in the GPR119
project we concentrated on gaining agonistic activity by taking
only the similarity to queries 1−4 into account and did not
consider any physicochemical properties for reagent selection.
It is also possible to enumerate virtual libraries according to the
reaction scheme and to select the most appropriate reagents
based on an assessment of the products, e.g., by looking at
docking or 3D alignment results. Last but not least we have
found it useful to manually enrich the in silico selected reagents
by making use of SAR information or other project specific
knowledge. In any case an extensive manual interaction driven
by the experience of the project team is usually included to
finalize the decoration of a core.
was already described in the literature and in patent
applications. Four structurally different known GPR119
agonists were selected as query molecules (see Figure
2).25−29 These compounds differ mainly in the central part
but display a similar left-hand and right-hand side decoration
pattern. They all share a methyl sulfone substituted aryl part on
the left-hand side as well as an alkyl carbamate on the other
side. On the basis of the hypothesis that these peripheral
moieties are important for the activity, these features were
chosen as “activity anchors”.
The FTrees fragment space search in BICLAIM provided
thousands of virtual hits per query. Usually each hit list has to
run through a postprocessing step on its own, but in this case
all query molecules overlapped well in 3D, although they were
coming from different structural classes.
In addition to standard BICLAIM campaigns in which
usually not much is known about the binding mode or SAR
contributions of the query molecules, we have also introduced
the so-called “activity anchor principle”. In our hands an activity
anchor is a structural element known to or at least supposed to
provide key contributions to a certain biological activity.
Usually the activity anchor is defined by analyzing the SAR of
the query ligands and literature or in-house information. The
presence or absence of an activity anchor or the option to
introduce the corresponding structural feature via the synthesis
protocol can be used to filter the hit lists for core selection
(step 4 in Figure 1). Furthermore the activity anchor could
itself serve as a core which (by applying suitable chemistry) can
be decorated by means of focused two-dimensional combina-
torial libraries. Alternatively, scaffold replacement tools like
CAVEAT,22 ReCore23 or the scaffold replacement and ligand
growing tools implemented in MOE24 can also be applied to
further elaborate activity anchors. This “activity anchor
principle” is widely used in Boehringer Ingelheim’s BICLAIM
campaigns. A recent analysis of several projects showed that
using an activity anchor doubles the chance of identifying true
hits in a BICLAIM virtual screening campaign.
This 3D overlap indicates that they probably all share an
identical binding epitope, which allowed us to combine all the
hits from different queries into one large hit list. For each query
molecule the most similar 10 000 FTrees hits were kept for
postprocessing.
In this case 3D alignments of the virtual hits were generated
with ROCS.29 As the respective query, we chose compound 2
because of its relatively high rigidity. The putative active 3D
conformation of 2 was defined based on the overlap with the
other three queries, as shown in Figure 2. The FTrees hits were
prepared for the ROCS alignments by applying Corina30and
Omega.31 Recently, the first pharmacophore model for
GPR119 agonists was published.32 However, at the time this
work was conducted, the pharmacophore model was not
available, and thus, this information could not be included in
our postprocessing analysis. The ROCS alignments were
visually inspected making use of the MOE mdb-browser
functionality24 to select the most promising scaffolds. In this
case the main focus was on molecules carrying both activity
anchors. From these molecules two cores were selected for
synthesis campaigns.
BICLAIM Virtual Screening and Synthesis Campaigns.
At the time at which the in-house virtual screening campaign
for GPR119 agonists was initiated, a wealth of different agonists
The first virtual hit class that was selected for synthesis in
combinatorial chemistry consisted of a 3,9-diaza-spiro[5.5]-
undecane core (see 5, Figure 3). Both activity anchors
11033
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
Figure 3. ROCS29 overlap of query 2 (green) with spiro compound 5 (yellow), general information on the core and its possible decoration as stored
in BICLAIM.
Scheme 1. Synthesis of Spiro Ureas 9a−c (Library 1)
Figure 4. ROCS29 overlap of query 2 (green) with spiro compound 10 (yellow), the general information on the core and its possible decoration as
stored in BICLAIM.
superimpose very well, whereas the spirocyclic core (yellow)
does not match optimally with the steric demand of the central
part of query 2 (green). However, we assumed that the core
only arranges the activity anchors correctly in space and
provides the right distance between the aryl and the carbamate
moiety and should itself not contribute too much to binding.
The suggested chemistry to address the left-hand side (LHS)
decoration involved reductive amination with aldehydes, amide
11034
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
Scheme 2. Synthesis of Spiro Dihydroisoxazoles 14a−c (Library 2)
Figure 5. Outcome and activity distribution of library 1.
bond formation with carboxylic acids, carbamate synthesis with
carbamoyl halides, and finally urea formations with isocyanates.
Proposed variations for the right-hand side (RHS) were
analogous; however, we concentrated on alkyl carbamates
that were similar to the query decoration (cf. the activity anchor
principle). The urea linker between core and aryl moiety
provided the best superimposition with the queries and was
therefore chosen for our initial synthesis campaign. According
to the activity anchor principle, we maintained a tert-butyl
carbamate on the right-hand side and varied the core as well as
the aryl part of the urea in a combinatorial library matrix. In
order to diversify the core with respect to orientation and
absolute length, not only 3,9-diaza-spiro[5.5]undecane core 8a
but also the cores 2,8-diaza-spiro[4.5]decane 8b and 2,7-diaza-
spiro[3.5]nonane 8c were employed (see Scheme 1). Because
of limited accessibility of isocyanates with the appropriate
decoration, we synthesized the targeted ureas using amines
from our in-house amine collection and a suitable coupling
reagent. During method development in order to optimize the
reaction conditions, we found that 1-1′-carbonyldi(1,2,4-
triazole) (7, CDT) was superior to other coupling reagents
like N,N′-carbonyldiimidazole (CDI) or 4-nitrophenyl chlor-
oformate.
The synthesis (Scheme 1) involved preactivation of 72
amines/anilines 6 with CDT and addition of the corresponding
readily available spiro compounds 8a−c on a 0.1 mmol scale.
1,4-Dioxane as a solvent and 1,8-diazabicyclo[5.4.0]undec-7-
ene (DBU) as a base showed the best conversions throughout
the differently reactive amines used. The crude products were
subsequently purified by automated preparative reversed phase
HPLC using narrow gradients. Yields were generally good,
varying between 20% and 90%, and a total number of 194 (out
of 216) compounds were submitted for in vitro testing for
GPR119 agonistic activity.
The second class of interesting virtual hits that was selected
for synthesis consisted of an alternative spirocyclic system (1-
oxa-2,8-diaza-spiro[4.5]dec-2-ene 10). The core, when properly
decorated, overlaps very well with the query molecule (see
11035
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
Figure 6. Outcome and activity distribution of library 2.
Figure 4). The spirocyclic system again allows the incorpo-
ration of both activity anchors and arranges them optimally in
the 3D space. Core 10 represents a very rigid system that we
expected not to lose too much conformational entropy upon
binding. Although the overlap appears obvious retrospectively,
the idea of using spirocyclic cores had so far not been in the
focus of our medicinal chemistry activities for GPR119 agonists.
Similar to the first library (vide supra) we decided to keep
the RHS carbamate moiety constant as an activity anchor while
varying the amines for the amide bond formation and
simultaneously also varying the core as depicted in Scheme 2.
The selection of amines was primarily based on an acceptable
superimposition with queries 1−4 and by taking into account
published and in-house SAR knowledge. Therefore, we focused
on primary (het)arylamines decorated with diverse substitu-
ents. The 6-ring spiro-dihydroisoxazole 11a (see Scheme 2)
gave proper superimpositions with the queries; however, its
pyrrolidine 5-ring analogue (see 11c) was also found
interesting. The 3-piperidino analogue of 11a, 11b was
additionally included in order to diversify the structural space.
Method development prior to library synthesis showed that
N,N,N′,N′-tetramethylchloroformamidinium hexafluorophos-
phate (TCFH, 12) gave the best conversions with respect to
the broad reactivity spectrum of the employed amines
compared to other coupling reagents like Mukaiyama’s reagent
(2-chloro-1-methylpyridinium iodide), TBTU, or its aza-
analogue HATU.33
were obtained in generally acceptable to good yields ranging
between 10% and 88%.
Biological Results and Hit Optimization. All synthesized
compounds were tested for their agonistic activity (potency as
EC50) and their efficacy (intrinsic activity, % IA) toward the
human GPR119 receptor using an ALPHAscreen cAMP assay.
In this assay, the intrinsic activity as stimulation of the receptor
is reported in %, the relative activity of the test compounds
being compared to the efficacy of N-(4-methanesulfonylphen-
yl)-5-nitro-6-[4-(phenylsulfanyl)piperidin-1-yl]pyrimidin-4-
amine,34 which was set to 200%. We regarded compounds
exceeding 160% IA stimulation as full agonists, whereas
compounds between 130% IA and 160% IA were considered
as partial agonists.
The outcome of the activity distribution of the initial in vitro
testing toward activity on human GPR119 receptors of
compounds derived from library 1 is depicted in Figure 5.
Overall we obtained disappointing results, as only ureas of a
3,9-diaza-spiro[5.5]undecane-type core (see 9a in library 1a)
yielded active compounds, albeit with only modest potency.
The most active derivative was 15, displaying an EC50 of 800
260 nM, a full agonist with an intrinsic activation of 165 9%
IA. However, the structure−activity relationships were rather
steep, which did not encourage us to invest in further syntheses.
Biological in vitro results for library 2 also showed a steep
SAR, again with a strong preference for one specific core
variation (see Figure 6). All synthesized 1-oxa-2,7-diaza-
spiro[4.5]dec-2-enes (see 14b in library 2b) or 1-oxa-2,7-
diaza-spiro[4.4]non-2-enes (see 14c in library 2c) were
inactive. However, we were pleased to discover that some
compounds of type 14a (see library 2a) displayed submicro-
molar potencies. The most active compound was 16 with an
EC50 of 93 48 nM showing full agonistic activity of 170
10% IA.
The library (three carboxylic acids 11a−c combined with 95
different alkyl- and (het)arylamines 13) was synthesized on a
0.1 mmol scale using TCFH (12) to preactivate the carboxylic
acid followed by addition of the corresponding amine 13 and
purification by means of automated preparative reversed phase
HPLC (narrow gradients). 261 of theoretically 285 compounds
11036
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
The vast differences in potency could not be explained solely
by the quality of the overlap compared to the queries, as all
molecules gave acceptable superimpositions. Especially when it
comes to small structural variations, computational methods
may not always sufficiently predict their influence. On the basis
of these promising results, we initiated a hit-to-lead approach in
order to further improve the potency and other in vitro
parameters (data not shown). Structural modifications involved
combinatorial variations of the LHS arylamine moiety which
were closely related to the difluoro-p-cyano substitution pattern
of 16. Also, we diversified the carbamate portion by introducing
further alkyl substituents but also included bioisosters and other
known binding motifs (see Figure 7).35
in Scheme 3 by TCFH mediated amide bond formation of 11a
with the corresponding aniline followed by boc-deprotection
with TFA yielding piperidine 17 which was then reacted with 5-
chloro-3-phenyl[1,2,4]oxadiazole. The alternative route to 18
via reaction of 17 with cyanogen bromide followed by ring
closure with N-hydroxybenzamidine resulted in a lower total
yield. Compared to 16, 18 exhibited increased agonistic activity
in terms of both potency (EC50 = 30 11 nM) and efficacy
(173 7% IA).
We then considered combinations of promising LHS and
RHS decorations, exemplified by spiro compound 21 (see
Scheme 4). HATU assisted amide coupling followed by
attachment of the phenyl substituted oxadiazole yielded 20
which was subsequently oxidized with mCPBA to give
methanesulfinyl substituted 21. 21 showed an EC50 of 21
15 nM, while being a borderline full agonist with an intrinsic
activity of 160 6% IA.
The structure−activity relationships obtained from 21, its
close analogues, and other related potent and fully efficacious
compounds with structurally diverse substitution patterns
(compounds and data not shown) were very promising.
Thus, we progressed 1-oxa-2,8-diaza-spiro[4.5]dec-2-ene type
compounds into a lead optimization program with the goal to
identify candidates for preclinical development in order to
provide new options to treat type 2 diabetes mellitus.
Figure 7. From hit to lead: strategies for structural modification of 16.
a
Scheme 3. Synthesis of Oxadiazole 18
SUMMARY AND CONCLUSION
■
The discovery of a novel lead class of GPR119 agonists was
achieved through a fruitful combination of virtual screening and
combinatorial chemistry using the BICLAIM approach. In our
group we successfully applied this technology in various
projects, yielding single hits, leads, or even advanced compound
classes. The pillars for this success are diverse. First, our vast
virtual BICLAIM space is based on numerous manifold
combinatorial libraries. The chemistry to synthesize the libraries
and their building blocks is well established in-house, allowing
for short cycle times from idea to synthesis. Second, it is
important to identify and select the most appropriate and
promising scaffolds followed by a careful selection of their
decoration. We noticed that the synthesis of virtual hits as
combinatorial library arrays using a much more diverse
decoration pattern than suggested by the single in silico hits
dramatically increases the probability of discovering real hits.
The design of these libraries is done as teamwork involving
computational, combinatorial, and medicinal chemists, making
extensive use of in-house and literature knowledge. In our
a
(a) TCFH, DIPEA, DCM, r.t.; (b) 4-amino-2,5-difluoro-benzonitrile;
(c) dichloromethane, TFA, 43% (over three steps); (d) 5-chloro-3-
phenyl-[1,2,4]oxadiazole, DIPEA, NMP, 160 °C, 50%.
4-(Pyrimidin-2-yl) substituents and 3-alkyl-1,2,4-oxadiazoles
attached to the nitrogen of the piperidine can be found in
recent patent literature as replacements for the carbamate
part.36,37 Therefore, compound 18 was synthesized as outlined
a
Scheme 4. Synthesis of Spiro-Compound 21
a
(a) HATU, DIPEA, dichloroethane, r.t.; (b) 2,5-difluoro-4-methylsulfanyl-phenyl amine; (c) HCl in diethyl ether, 44% (over three steps); (d) 5-
chloro-3-phenyl-[1,2,4]oxadiazole, DIPEA, NMP, 150 °C, 51%; (e) 3-chloroperoxybenzoic acid, dichloromethane, 68%.
11037
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
NaCl, 3.6 mM KCl, 0.5 mM NaH2PO4, 0.5 mM MgSO4, 1.5 mM
CaCl2, 10 mM Hepes, 5 mM NaHCO3, pH 7.4, 0.5 mM IBMX, and
0.1% BSA; the final DMSO concentration is 1%). After 45 min of
incubation at room temperature (approximately 20 °C), the cAMP
concentrations are determined using the ALPHAscreen cAMP assay
kit (catalog no. 6760625R from PerkinElmer). An amount of 10 μL of
biotin-cAMP-bead solution (final concentration of 1 U/well in lysing
buffer (5 mM Hepes (pH 7.4), 0.1% BSA, 0.5% Tween)) is added.
The plates are incubated for another 2 h at room temperature. The
cAMP concentrations are calculated using a cAMP standard curve
from the Alpha screen counts. The data analysis is carried out by
calculating the EC50 value and the maximum value based on a positive
control (N-(4-methanesulfonylphenyl)-5-nitro-6-[4-(phenylsulfanyl)-
piperidin-1-yl]pyrimidin-4-amine),32 using suitable software (XLfit).
By default six measurements were conducted and the EC50 was
calculated using the geometric mean of the individual results
(arithmetic mean for intrinsic activation).
experience, the synthesis only of prototypic virtual hits as
diagnostic compounds has turned out to be inferior. Third, the
concept of an “activity anchor” has been proven rewarding and
also increases the chances of finding true hits. We successfully
applied the above-mentioned factors for the in silico discovery,
selection, and design of two combinatorial libraries, which
yielded two new structural classes of potent GPR119 agonists.
Subsequent improvement of spiro compound 16 finally led to a
new lead class.
EXPERIMENTAL SECTION
Virtual Screening. The query molecules were generated in ISIS
Draw38 and exported as mol files. Hydrogens and formal charges were
added with Corina:30
■
corina −i t=sdf −o t=mol2 -d wh −o fcharges input.mol
query.mol2
For the fragment space search with the FTrees software17 the
following script was used:
Synthesis. All commercially available reagents were used without
further purification unless otherwise stated. 3,9-Diaza-spiro[5.5]-
undecane-3-carboxylic acid tert-butyl ester (8a, MFCD05861627)
was purchased from Small Molecules, Inc., Hoboken, NJ, U.S. tert-
Butyl 2,8-diaza-spiro[4.5]decane-8-carboxylate (8b, MFCD09608078),
tert-butyl 2,7-diaza-spiro[3.5]nonane-2-carboxylate (8c,
MFCD09834803), 8-(tert-butoxycarbonyl)-1-oxa-2,8-diaza-spiro[4.5]-
dec-2-ene-3-carboxylic acid (11a, MFCD12198539), 7-(tert-butoxy-
carbonyl)-1-oxa-2,7-diaza-spiro[4.5]dec-2-ene-3-carboxylic acid (11b,
MFCD12198540), and 7-(tert-butoxycarbonyl)-1-oxa-2,7-diaza-spiro-
[4.4]non-2-ene-3-carboxylic acid (11c, MFCD12198541) were
1
purchased from WuXi App Tec Co. Ltd., Shanghai, China. H NMR
spectra were recorded on Bruker DRX 400 and DPX 400
spectrometers. Chemical shifts are reported in ppm using the residual
of deuterated solvents as an internal standard. FTIR spectra were
recorded using a Thermo Nicolet Avatar system 370. Prior to
biological testing, NMR and/or analytical reversed phase HPLC was
used to verify the purity of the compounds. All tested compounds
derived from combinatorial libraries were at least 80% pure; all other
compounds including resynthesis of the most interesting hits were at
least 95% pure.
2,5-Difluoro-4-methylsulfanylphenylamine. 4-Bromo-2,5-di-
fluorophenylamine (0.91 g, 4.4 mmol), CuI (120 mg, 0.63 mmol),
NaI (1.31 g, 8.7 mmol), and dimethylethylenediamine (0.12 mL, 0.9
mmol) in 1,4-dioxane (4.0 mL) were purged with argon, and the
mixture was stirred in a pressure pipe at 140 °C for 3 h. After the
mixture was cooled to room temperature DCM was added, and the
mixture was then washed consecutively with aqueous HCl solution (1
mol/L), aqueous Na2CO3 solution (10%), and brine. The organic
layer was concentrated in vacuo to yield 2,5-difluoro-4-iodophenyl-
amine (790 mg, 71%) which was used without further purification.
Under an argon atmosphere 2,5-difluoro-4-iodophenylamine (790 mg,
3.1 mmol) was dissolved in dry DMF (4.0 mL), and dimethyl disulfide
(0.68 mL, 7.7 mmol) was added followed by addition of NiBr2 (80 mg,
0.37 mmol), 2,2-dipyridyl (50 mg, 0.32 mmol), and zinc dust (410 mg,
6.3 mmol). The reaction mixture was allowed to react at 80 °C for 2 h.
After the mixture was cooled to room temperature DCM and active
charcoal were added and the mixture was filtered through powdered
cellulose. The remaining liquid is consecutively washed with aqueous
Na2CO3 solution (10%) and brine. The organic layer was concentrated
in vacuo to afford 2,5-difluoro-4-methylsulfanylphenylamine (400 mg,
74%) which was used without further purification. C7H7F2NS (M =
175.2 g/mol). ESI-MS: 176 [M + H]+. tR (HPLC): 1.09 min (Waters
ZQ MS, Alliance 2690/2695, Waters XBridge C18 3.5 μm, 4.6 mm ×
20 mm IS, flow 4.0 mL/min, temperature 40 °C, H2O (+0.1% TFA)/
MeOH).
The smiles files were converted to SD files, and the structures
were unified using Pipeline Pilot.39
For the postprocessing with ROCS29 the following commands were
used:
corina −i t=sdf −o t=sdf −d wh hits.sdf hits_3D.sdf
omega −in hits_3D.sdf −out hits_conf.oeb
rocs −query query.mol2 −dbase hits_conf.oeb −prefix
hits_aligned
The SD files resulting from ROCS29 were imported in a MOE
database and visually inspected with the database browser.24
Determination of Agonistic Activity (hGPR119 Assay). The
effect of the compounds on the activation of GPR119 and on the
stimulation of intracellular cAMP concentration is determined using
the ALPHAscreen cAMP assay kit (catalog no. 6760625R) made by
PerkinElmer. MIN6 cells40 are stably transfected with an expression
vector for human GPR119 cDNA (accession no. NP_848566). Min-
6/hGPR119 cells are cultured in DMEM, 10% FBS, 50 μM β-
mercaptoethanol, 0.6 mg/mL geneticin, 2 mM GlutaMAX at 37 °C,
5% CO2. For the assay, the cells are seeded in culture plates (white,
384-well, TC, sterile with lid, catalog no. 6006280 (Perkin-Elmer);
10000 cells/well; 50 μL). The plates covered with lids are then
incubated for 24 h at 37 °C/5% CO2. After the medium is aspirated
from the wells completely, an amount of 10 μL of the test compounds
is added, the compounds are diluted using stimulating buffer (140 mM
General Procedure for the Synthesis of Spiro Ureas 9a−c
(Library 1). To a stirred solution of amine 6 (0.10 mmol) in 1,4-
dioxane (2.5 mL) were added sequentially DBU (30 μL, 0.20 mmol)
and CDT (7, 33 mg, 0.20 mmol). The mixture was stirred for 5 min at
room temperature prior to addition of spiro-amine 8 (0.11 mmol) in
1,4-dioxane (1.0 mL). The reaction mixture was allowed to react at
room temperature for 12 h. Subsequently the mixture was
11038
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
concentrated in vacuo and the residue was purified by preparative
reversed phase HPLC (C18 RP Sunfire column, gradient MeOH/H2O
(+0.1% TFA)) to afford the desired urea derivative.
purification. C15H14F2N4O2 (M = 320.3 g/mol). ESI-MS: 321 [M +
H]+. tR (HPLC): 0.93 min (Waters ZQ MS, Alliance 2690/2695,
Waters XBridge C18 3.5 μm, 4.6 mm × 20 mm IS, flow 4.0 mL/min,
temperature 40 °C, H2O (+0.1% TFA)/MeOH). IR (ATR) vmax 3344,
9-(2-Bromo-4-methanesulfonylphenylcarbamoyl)-3,9-diaza-
spiro[5.5]undecane-3-carboxylic Acid tert-Butyl Ester (15).
Compound 15 was synthesized following the general procedure
described above. 3,9-Diaza-spiro[5.5]undecane-3-carboxylic acid tert-
butyl ester (8a) and 2-bromo-4-methanesulfonylphenylamine were
used as starting materials. Yield: 21.5 mg (41%). C22H32BrN3O3S (M
= 530.5 g/mol). ESI-MS: 530/532 [M + H]+. tR (HPLC): 2.05 min
(Agilent 1100 with DA and Waters MS detector, Sunfire C18, 4.6 mm
× 50 mm, 3.5 μm (Waters), flow 1.5 mL/min, temperature 60 °C,
H2O (+0.1% TFA)/acetonitrile (+0.08% TFA)). 1H NMR (400 MHz,
DMSO-d6): δ 1.38−1.43 (m, 4H), 1.39 (s, 9H), 1.45−1.50 (m, 4H),
3.24 (s, 3 H), 3.26−3.34 (m, 4H), 3.43−3.48 (m, 4H), 7.83 (dd, 1H, J
= 8.6, 2.0 Hz), 7.88 (d, 1H, J = 8.6 Hz), 8.08 (d, 1H, J = 2.0 Hz), 8.24
(s, 1H).
General Procedure for the Synthesis of Spiro Amides 14a−c
(Library 2). To a stirred solution of acid 11 (0.10 mmol) in THF (1.0
mL) were added sequentially DIPEA (52 μL, 0.30 mmol) and chloro-
N,N,N′,N′-tetramethylformamidinium hexafluorophosphate (TCFH
12, 45 mg, 0.16 mmol). The mixture was stirred for 45 min at
room temperature prior to addition of a solution of the corresponding
amine 13 (0.12 mmol) in THF (1.0 mL). The reaction mixture was
allowed to react at room temperature for 12 h. Subsequently DMF was
added and the mixture was filtered through basic aluminum oxide
followed by washing with DMF/MeOH (9:1). The mixture was
concentrated in vacuo, and the residue was taken up in DMF and was
purified by preparative reversed phase HPLC (C18 RP Sunfire
column, gradient MeOH/H2O (+0.1% TFA)) to afford the desired
amide.
3-(4-Cyano-2,5-difluorophenylcarbamoyl)-1-oxa-2,8-diaza-
spiro[4.5]dec-2-ene-8-carboxylic Acid tert-Butyl Ester (16). To
a stirred solution of 8-(tert-butoxycarbonyl)-1-oxa-2,8-diaza-spiro[4.5]-
dec-2-ene-3-carboxylic acid (11a, 150 mg, 0.53 mmol) in THF (3.0
mL) were added sequentially DIPEA (0.27 mL, 1.58 mmol) and
chloro-N,N,N′,N′-tetramethylformamidinium hexafluorophosphate
(TCFH 12, 237 mg, 0.84 mmol). The mixture was stirred for 45
min at room temperature prior to addition of 4-amino-2,5-
difluorobenzonitrile (89 mg, 0.60 mmol). The reaction mixture was
allowed to react at room temperature for 12 h. The reaction mixture
was quenched with water and was extracted with EtOAc. The
combined organic layers were concentrated in vacuo. The residue was
taken up in DMF and was purified by preparative reversed phase
HPLC (C18 RP Sunfire column, gradient MeOH/H2O (+0.1% TFA))
to afford compound 16 (73 mg, 33%). C20H22F2N4O4 (M = 420.4 g/
mol). ESI-MS: 421.2 [M + H]+. tR (HPLC): 0.83 min (Waters Acquity
with DA and MS detector, Sunfire C18, 2.1 mm × 50 mm, 2.5 μm
(Waters), flow 1.5 mL/min, temperature 60 °C, H2O (+0.1% formic
acid)/acetonitrile (+0.1% formic acid)). 1H NMR (400 MHz, DMSO-
d6): δ 1.41 (s, 9H), 1.71−1.77 (m, 4H), 3.11 (s, 2H), 3.35−3.47 (m, 4
H), 7.88−7.95 (m, 1H), 8.04−8.09 (m, 1H), 10.41 (s, 1H).
1-Oxa-2,8-diaza-spiro[4.5]dec-2-ene-3-carboxylic Acid (4-
Cyano-2,5-difluorophenyl)amide (17). To a stirred solution of
8-(tert-butoxycarbonyl)-1-oxa-2,8-diaza-spiro[4.5]dec-2-ene-3-carbox-
ylic acid (11a, 250 mg, 0.88 mmol) in DCM (40 mL) were added
sequentially DIPEA (0.31 mL, 1.80 mmol) and chloro-N,N,N′,N′-
tetramethylformamidinium hexafluorophosphate (310 mg, 1.10
mmol). Subsequently 4-amino-2,5-difluorobenzonitrile (170 mg, 1.10
mmol) was added and the reaction mixture was allowed to react at
room temperature for 7 days. The mixture was then washed
consecutively with aqueous HCl solution (1 mol/L), aqueous
Na2CO3 solution, and brine. The organic layer was treated with
TFA (2 mL), and the mixture was stirred at room temperature for 2 h.
The mixture was diluted with DCM, and aqueous HCl solution (50
mL, 1 mol/L) was added. To the aqueous layer was added DCM, and
aqueous Na2CO3 solution was added until the mixture was basic. The
organic layer was washed with brine followed by drying over Na2SO4
and concentration in vacuo to yield compound 17 (120 mg, 43%) as
the corresponding hydrochloride, which was used without further
1
2236, 1696, 1526, 1485, 1430, 1237, 909, 643 cm−1. H NMR (400
MHz, DMSO-d6): δ 1.65−1.72 (m, 4H), 2.58−2.69 (m, 2H), 2.85−
2.93 (m, 2H), 3.05 (s, 2H), 7.92−8.02 (m, 2H).
8-(3-Phenyl[1,2,4]oxadiazol-5-yl)-1-oxa-2,8-diaza-spiro[4.5]-
dec-2-ene-3-carboxylic Acid (4-Cyano-2,5-difluorophenyl)-
amide (18). 17 (18 mg, 0.056 mmol) was dissolved in NMP (1.0
mL), and DIPEA (40 μL, 0.23 mmol) and 5-chloro-3-phenyl-1,2,4-
oxadiazole (13 mg, 0.072 mmol) were added. The reaction mixture
was allowed to react at 160 °C for 15 min. After the mixture was
cooled to room temperature DCM was added and the mixture was
washed consecutively with aqueous HCl solution (2 mol/L), aqueous
Na2CO3 solution (10%), and brine. The organic layer was
concentrated in vacuo, and the residue was purified by preparative
MPLC (silica, DCM/MeOH = 98:2). The product containing
fractions were concentrated in vacuo, and the residue was crystallized
from diethyl ether/petrol ether to afford compound 18 (13 mg, 50%)
as a colorless solid. C23H18F2N6O3 (M = 464.4 g/mol). ESI-MS: 465
[M + H]+. tR (HPLC): 1.52 min (Waters ZQ MS, Alliance 2690/2695,
Waters XBridge C18 3.5 μm, 4.6 mm × 20 mm IS, flow 4.0 mL/min,
temperature 40 °C, H2O (+0.1% TFA)/MeOH). 1H NMR (400 MHz,
DMSO-d6): δ 1.93−1.98 (m, 4H), 3.19 (s, 2H), 3.67−3.80 (m, 4H),
7.48−7.55 (m, 3H), 7.89−7.96 (m, 3H), 8.04−8.10 (m, 1H), 10.46 (s,
1H).
1-Oxa-2,8-diaza-spiro[4.5]dec-2-ene-3-carboxylic Acid (2,5-
Difluoro-4-methylsulfanylphenyl)amide (19). To a stirred
solution of 8-(tert-butoxycarbonyl)-1-oxa-2,8-diaza-spiro[4.5]dec-2-
ene-3-carboxylic acid (11a, 600 mg, 2.11 mmol) in 1,2-dichloroethane
(20 mL) were added sequentially DIPEA (0.65 mL, 3.780 mmol) and
HATU (1.2 g, 3.2 mmol). The mixture was stirred at room
temperature for 45 min. Subsequently 2,5-difluoro-4-methylsulfanyl-
phenylamine (400 mg, 2.28 mmol) was added and the reaction
mixture was allowed to react at room temperature for 3 h. The mixture
was then washed consecutively with aqueous HCl solution (1 mol/L),
aqueous Na2CO3 solution, and brine. The organic layer was
concentrated and taken up in ethanol/DCM. In order to remove
the boc protection group, the mixture was treated with etheric HCl
and the mixture was stirred at room temperature for 2 h. After
completion of the reaction the precipitate was collected by filtration
and was dried in vacuo to yield compound 19 (350 mg, 44%) as the
corresponding HCl salt. C15H17F2N3O2S·HCl (M = 377.8 g/mol).
ESI-MS: 342 [M + H]+. tR (HPLC): 1.13 min (Waters ZQ MS,
Alliance 2690/2695, Waters XBridge C18 3.5 μm, 4.6 mm × 20 mm
IS, flow 4.0 mL/min, temperature 40 °C, H2O (+0.1% TFA)/MeOH).
1H NMR (400 MHz, DMSO-d6): δ 1.98−2.04 (m, 4H), 2.50 (s, 3H),
3.13−3.19 (m, 4H), 3.27 (s, 2H), 7.29−7.35 (m, 1H), 7.43−7.49 (m,
1H), 8.94 (bs, 2H), 10.16 (s, 1H).
8-(3-Phenyl[1,2,4]oxadiazol-5-yl)-1-oxa-2,8-diaza-spiro[4.5]-
d e c - 2 - e n e - 3 - c a r b o x y l i c A c i d ( 2 , 5 - D i fl u o r o - 4 -
methylsulfanylphenyl)amide (20). To a stirred solution of 19
(350 mg, 0.93 mmol) in NMP (5.0 mL) were added sequentially
DIPEA (0.65 mL, 3.76 mmol) and 5-chloro-3-phenyl[1,2,4]oxadiazole
(250 mg, 1.4 mmol). The mixture was stirred for 10 min at 150 °C.
After cooling to room temperature, the reaction mixture was washed
consecutively with aqueous HCl solution (2 mol/L), aqueous Na2CO3
solution (10%), and brine. The organic layer was concentrated under
reduced pressure, and the residue was purified by MPLC (silica,
DCM/MeOH = 97:3). The product containing fractions were
concentrated in vacuo and the residue was crystallized from diethyl
ether to afford compound 20 (230 mg, 51%) as a colorless solid.
C23H21F2N5O3S (M = 485.5 g/mol). ESI-MS: 486 [M + H]+. tR
(HPLC): 1.59 min (Waters ZQ MS, Alliance 2690/2695, Waters
XBridge C18 3.5 μm, 4.6 mm × 20 mm IS, flow 4.0 mL/min,
temperature 40 °C, H2O (+0.1% TFA)/MeOH). 1H NMR (400 MHz,
DMSO-d6): δ 1.92−1.98 (m, 4H), 2.51 (s, 3H), 3.17 (s, 2H), 3.71−
3.77 (m, 4H), 7.30−7.36 (m, 1H), 7.45−7.55 (m, 4H), 7.89−7.93 (m,
2H), 10.16 (s, 1H).
11039
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
8-(3-Phenyl[1,2,4]oxadiazol-5-yl)-1-oxa-2,8-diaza-spiro[4.5]-
d e c - 2 - e n e - 3 - c a r b o x y l i c A c i d ( 2 , 5 - D i fl u o r o - 4 -
methanesulfinylphenyl)amide (21). 20 (100 mg, 0.21 mmol)
was dissolved in DCM (10 mL), and the mixture was cooled to 0 °C.
3-Chloroperoxybenzoic acid (38 mg, 0.22 mmol) was added, and the
reaction mixture was allowed to react at 0 °C for 2 h. The mixture was
then extracted with aqueous NaOH solution (1 mol/L), and the
organic layer was washed with brine followed by drying over Na2SO4
and concentration in vacuo. The residue was triturated with diethyl
ether, filtered, and dried to yield compound 21 (70 mg, 68%) as a
colorless solid. C23H21F2N5O4S (M = 501.5 g/mol). ESI-MS: 502 [M
+ H]+. tR (HPLC): 1.44 min (Waters Alliance with DA and MS
detector, Waters XBridge C18 3.5 μm, 4.6 mm × 30 mm, flow 4.0
mL/min, temperature 60 °C, H2O (+0.1% TFA)/MeOH (+0.1%
(4) Jones, R. M.; Leonard, J. N.; Buzard, D. J.; Lehmann, J. GPR119
agonists for the treatment of type 2 diabetes. Expert Opin. Ther. Pat.
2009, 19, 1339−1359.
(5) Chu, Z.-L.; Carroll, C.; Chen, R.; Alfonso, J.; Gutierrez, V.; He,
H.; Lucmn, A.; Xing, C.; Sebring, K.; Zhou, J.; Wagner, B.; Unett, D.;
Jones, R. M.; Behan, D. P.; Leonard, J. N-Oleoyldopamine enhances
glucose homeostasis through the activation of GPR119. Mol.
Endocrinol. 2010, 24, 161−170.
(6) Ezzili, C.; Otrubova, K.; Boger, D. L. Fatty acid amide signaling
molecules. Bioorg. Med. Chem. 2010, 20, 5959−5968.
(7) Ohishi, T.; Yoshida, S. The therapeutic potential of GPR119
agonists for type 2 diabetes. Expert Opin. Invest. Drugs 2012, 21, 321−
328.
(8) Semple, G.; Fioravant, B.; Pereira, G.; Calderon, I.; Uy, J.; Choi,
K.; Xiong, Y.; Ren, A.; Morgan, M.; Dave, V.; Thomsen, W.; Unett, D.
J.; Xing, C.; Bossie, S.; Carroll, C.; Chu, Z.-L.; Grottick, A. J.; Hauser,
E. K.; Leonard, J.; Jones, R. M. Discovery of the first potent and orally
efficacious agonist of the orphan G-protein coupled receptor 119. J.
Med. Chem. 2008, 51, 5172−5175.
(9) Wu, Y.; Kuntz, J. D.; Carpenter, A. J.; Fang, J.; Sauls, H. R.;
Gomez, D. J.; Ammala, C.; Xu, Y.; Hart, S.; Tadepalli, S. 2,5-
Disubstituted pyridines as potent GPR 119 agonists. Bioorg. Med.
Chem. Lett. 2010, 20, 2577−2581.
(10) McClure, K. F.; Darout, E.; Guinaraes, C. R. W.; DeNinno, M.
P.; Mascitti, V.; Munchhof, M. J.; Robinson, R. P.; Kohrt, J.; Harris, A.
R.; Moore, D. E.; Li, B.; Samp, L.; Lefker, B. A.; Futatsugi, K.; Kung,
D.; Bonon, P. D.; Cornelius, P.; Wang, R.; Salter, E.; Hornby, S.;
Kalgutkar, A. S.; Chen, Y. Activation of the G-protein-coupled receptor
119: a conformation-based hypothesis for understanding agonist
response. J. Med. Chem. 2011, 54, 1948−1952.
(11) Semple, G.; Lehmann, J.; Wong, A.; Ren, A.; Bruce, M.; Shin, Y.-
J.; Sage, C. R.; Morgan, M.; Chen, W.-C.; Sebring, K.; Chu, Z.-L.;
Leonard, J.; A-Shamma, H.; Grottick, A. J; Du, F.; Liang, Y.; Demarest,
K.; Jones, R. M. Discovery of a second generation agonist of the
orphan G-protein coupled receptor GPR119 with an improved profile.
Bioorg. Med. Chem. Lett. 2012, 22, 1750−1755.
1
TFA)). H NMR (400 MHz, DMSO-d6): δ 1.94−1.99 (m, 4H), 2.87
(s, 3H), 3.19 (s, 2H), 3.71−3.78 (m, 4H), 7.49−7.55 (m, 3H), 7.56−
7.61 (m, 1H), 7.72−7.77 (m, 1H), 7.89−7.93 (m, 2H), 10.38 (s, 1H).
AUTHOR INFORMATION
Corresponding Author
■
*Phone: +49-(0)7351-5494701. Fax: +49-(0)7351-8394701. E-
mail: bernd.nosse@boehringer-ingelheim.com.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Georg Dahmann for assistance in the
preparation of this manuscript. Additionally we thank Steffen
Breitfelder, Peter Eickelmann, and Remko A. Bakker for
valuable discussions and Angelika Raible-Poschl as well as
̈
Esther Schmidt for biological activity measurements.
ABBREVIATIONS USED
■
cAMP, cyclic adenosine monophosphate; BICLAIM, Boeh-
ringer Ingelheim comprehensive library of accessible innovative
molecules; CDT, 1,1′-carbonyldi(1,2,4-triazole); DBU, 1,8-
diazabicyclo[5.4.0]undec-7-ene; DCM, dichloromethane;
DIPEA, N,N-diisopropylethylamine; FBS, fragment based
screening; FTrees, feature trees; GLP-1, glucagon-like peptide
1; GPR119, G-protein-coupled receptor 119; HATU, 2-(7-aza-
1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluoro-
phosphate; HTS, high-throughput screening; LHS, left-hand
side; NMP, N-methyl-2-pyrrolidinone; RHS, right-hand side; rt,
room temperature; SAR, structure−activity-relationship;
TCFH, chloro-N,N,N′,N′-tetramethylformamidinium hexa-
fluorophosphate; TFA, trifluoroacetic acid; THF, tetrahydro-
furan; VS, virtual screening
(12) Negoro, K.; Yonetoku, Y.; Maruyama, T.; Yoshida, S.; Takeuchi,
M.; Ohta, M. Synthesis and structure−activity relationship of 4-amino-
2-phenylpyrimidine derivatives as a series of novel GPR119 agonists.
Bioorg. Med. Chem. 2012, 20, 2369−2375.
(13) Scott, J. S.; Birch, A. M.; Brocklehurst, K. J.; Broo, A.; Brown, H.
̈
S.; Butlin, R. J.; Clarke, D. S.; Davidsson, O.; Ertan, A.; Goldberg, K.;
Groombridge, S. D.; Hudson, J. A.; Laber, D.; Leach, A. G.; MacFaul,
P. A.; McKerrecher, D.; Pickup, A.; Schofield, P.; Svensson, P. H.;
Sorme, P.; Teague, J. Use of small-molecule crystal structures to
̈
address solubility in a novel series of G protein coupled receptor 119
agonists: optimization of a lead and in vivo evaluation. J. Med. Chem.
2012, 55, 5361−5379.
(14) Lessel, U.; Wellenzohn, B.; Lilienthal, M.; Claussen, H.
Searching fragment spaces with feature trees. J. Chem. Inf. Model.
2009, 49, 270−279.
(15) Boehm, M.; Wu, T.-Y.; Claussen, H.; Lemmen, C. Similarity
searching and scaffold hopping in synthetically accessible combinato-
rial chemistry spaces. J. Med. Chem. 2008, 51, 2468−2480.
(16) Rarey, M.; Dixon, J. S. Feature trees: a new similarity measure
based on tree matching. J. Comput.-Aided Mol. Des. 1998, 12, 471−490.
(17) FTREES, version 2.3.0; Biosolve IT GmbH; St. Augustine, FL,
2010.
(18) Rarey, M.; Stahl, M. Similarity searching in large combinatorial
chemistry spaces. J. Comput.-Aided Mol. Des. 2001, 15, 497−520.
(19) Fischer, J. R.; Lessel, U.; Rarey, M. LoFT: Similarity-driven
multiobjective focused library design. J. Chem. Inf. Model. 2009, 50, 1−
21.
(20) Fischer, J. R.; Lessel, U.; Rarey, M. Improving similarity-driven
library design: customized matching and regio-selective feature trees. J.
Chem. Inf. Model. 2011, 51, 2156−2163.
(21) Lessel, U.; Wellenzohn, B.; Fischer, J. R.; Rarey, M. Design of
combinatorial libraries for the exploration of virtual hits from fragment
space searches with LoFT. J. Chem. Inf. Model. 2012, 52, 373−379.
REFERENCES
■
(1) (a) Wild, S.; Roglic, G.; Green, A.; Sicree, R.; King, H. Global
prevalence of diabetes: estimates for the year 2000 and projections for
2030. Diabetes Care 2004, 27, 1047−1053. (b) Diabetes Atlas, 2nd ed.;
International Diabetes Federation: Brussels, Belgium, 2003. (c)
International Diabetes Federation Diabetes Atlas. The Global Burden.
http://www.idf.org/diabetesatlas/5e/the-global-burden (accessed Sep-
tember 2012).
(2) Meece, J. Pancreatic islet dysfunction in type 2 diabetes: a
rational target for incretin-based therapies. Curr. Med. Res. Opin. 2007,
23, 933−944.
(3) Overton, H. A.; Babbs, A. J.; Doel, S. M.; Fyfe, M. C.; Gardner, L.
S.; Griffin, G.; Jackson, H. C.; Procter, M. J.; Rasamison, C. M.; Tang-
Christensen, M.; Widdowson, P. S.; Williams, G. M.; Reynet, C.
Deorphanization of a G protein-coupled receptor for oleoylethanola-
mide and its use in the discovery of small-molecule hypophagic agents.
Cell Metab. 2006, 3, 167−175.
11040
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041

Journal of Medicinal Chemistry
Article
(22) Lauri, G.; Bartlett, P. A. CAVEAT: a program to facilitate the
design of organic molecules. J. Comput.-Aided Mol. Des. 1994, 8, 51−
66.
(23) Maass, P.; Schulz-Gasch, T.; Stahl, M.; Rarey, M. ReCore: a fast
and versatile method for scaffold hopping based on small molecule
crystal structure conformations. J. Chem. Inf. Model. 2007, 47, 390−
399.
(24) Molecular Operating Environment (MOE), version 2010.10;
Chemical Computing Group Inc. (1010 Sherbooke St. West, Suite No.
910, Montreal, Quebec, Canada, H3A 2R7), 2010.
(25) Jones, R. M.; Semple, G.; Xiong, Y.; Shin, Y.-J.; Ren, A. S.;
Lehmann, J.; Fioravanti, B.; Bruce, M. A.; Choi, J., Sun, K. Substituted
Aryl and Heteroaryl Derivatives as Modulators of Metabolism and the
Prophylaxis and Treatment of Disorders Related Thereto. WO-
2005121121, 2005; Arena Pharmaceuticals, Inc.
Withall, D. M. Cycloamino Derivatives as GPR119 Agonists. WO-
2011147951, 2011; Prosidion Ltd.
(38) MDL ISIS/Draw, version 2.5 SR 3; MDL Information Systems,
Inc.: San Leandro, CA, 1990−2003.
(39) Pipeline Pilot, version 8.0.1; Accelrys Software Inc: San Diego,
CA, 2010.
(40) Miyazaki, J.; Araki, K.; Yamato, E.; Ikegami, H.; Asano, T.;
Shibasaki, Y.; Oka, Y.; Yamamura, K. Establishment of a pancreatic
beta cell line that retains glucose-inducible insulin secretion: special
reference to expression of glucose transporter isoforms. Endocrinology
1990, 127, 126−132.
(26) Fevig, J. M.; Wacker, D. A. [6,6] and [6,7]-Bicyclic GPR119 G
Protein-Coupled Receptor Agonists. WO-2008137435, 2008; Bristol-
Myers Squibb Company.
(27) Jones, R. M.; Lehmann, J. Modulators of Metabolism and the
Treatment of Disorders Related Thereto. WO-2008005576, 2008;
Arena Pharmaceuticals Inc.
(28) Bradley, S. E.; Fyfe, M. C. T.; Bertram, L. S.; Gattrell, W.;
Jeevaratnam, R. P.; Keily, J.; Procter, M. J.; Rasamison, C. M.;
Rushworth, P. J.; Sambrook-Smith, C. P.; Stonehouse, D. F.; Swain, S.
A.; Williams, G. M. GPCR Agonists. WO-2007003962, 2007;
Prosidion Ltd.
(29) ROCS, version 3.1.0; OpenEye Scientific Software, Inc.: Santa
Fe, NM, 2010.
(30) Corina, version 3.46; Molecular Networks GmbH: Erlangen,
Germany, 2008.
(31) Omega, version 2.4.3; OpenEye Scientific Software, Inc.: Santa
Fe, NM, 2010.
(32) Zhu, X.; Huang, D.; Lan, X.; Tang, C.; Zhu, Y.; Han, J.; Huang,
W.; Qian, H. The first pharmacophore model for potent G protein-
coupled receptor 119 agonist. Eur. J. Med. Chem. 2011, 46, 2901−
2907.
(33) Valeur, E.; Bradley, M. Amide bond formation: beyond the myth
of coupling reagents. Chem. Soc. Rev. 2009, 38, 606−631.
(34) Jones, R. M.; Semple, G.; Fioravanti, B.; Pereira, G.; Calderon,
I.; Uy, J.; Duvvuri, K.; Choi, J. S. K.; Xiong, Y.; Dave, V. 1,2,3-
Trisubstituted Aryl and Heteroaryl Derivatives as Modulators of
Metabolism and the Prophylaxis and Treatment of Disorders Related
Thereto Such as Diabetes and Hyperglycemia. WO-2004065380,
2004; Arena Pharmaceuticals Inc.
(35) Buzard, D. J.; Lehmann, J.; Han, S.; Jones, R. M. GPR119
agonists 2009−2011. Pharm. Pat. Anal. 2012, 1, 285−299.
(36) (a) Bertram, L. S.; Fyfe, M. C. T.; Jeevaratnam, R. P.; Keily, J.;
Krulle, T. M.; Rasamison, C. M.; Sambrook-Smith, C. P.; Swain, S. A.
Piperidinyl GPCR Agonists. WO-2010004343, 2010; Prosidion
Limited. (b) Wood, H. B.; Szewczyk, J. W.; Huang, Y.; Adams, A.
D. Substituted Cyclopropyl Compounds, Compositions Containing
Such Compounds and Methods of Treatment. WO-2009129036,
2009; Merck & Co. Inc. (c) Wood, H. B.; Adams, A. D.; Szewczyk, J.
W.; Zhang, Y.; Yang, M. Substituted Cyclopropyl Compounds,
Compositions Containing Such Compounds and Methods of
Treatment. WO-2011019538, 2011; Merck Sharp & Dohme Corp.
(d) Birch, A. M; Broo, D. A.; Butlin, R. J.; Clarke, D. S.; Davidsson, O.
P.; De La Motte, H.; Johansson, K. E.; Leach, A.; Macfaul, P. A.;
O’Donnell, C. J.; Scott, J. S.; Whittamore, P. R. O. 4-(Pyrimidin-2-yl)-
piperazine and 4-(Pyrimidin-2-yl)-piperidine Derivatives as GPR119
Modulators. WO-2011030139, 2011; AstraZeneca UK Limited.
(37) (a) Bertram, L. S.; Fyfe, M. C. T. Piperidinyl GPCR Agonists.
WO-2010004345, 2010; Prosidion Limited. (b) Carpenter, A. J.; Fang,
J.; Peckham, G. Chemical Compounds and Uses. WO-2010014593,
2010; Smithkline Beecham Corporation. (c) Barba, O.; Bell, J. C.;
Dupree, T. B.; Fry, P. T.; Bertram, L. S.; Fyfe, M. C. T.; Gattrell, W.;
Jeevaratnam, R. P.; Keily, J.; Krulle, T. M.; Mcdonald, R. W.; Morgan,
T.; Rasamison, C. M.; Schofield, K. L.; Stewart, A. J. W.; Swain, S. A.;
11041
dx.doi.org/10.1021/jm301549a | J. Med. Chem. 2012, 55, 11031−11041