Development of a plate-based optical biosensor fragment screening methodology to identify phosphodiesterase 10A inhibitors

Article abstract of DOI:10.1021/jm301665y

We describe the development of a novel fragment screening methodology employing a plate-based optical biosensor system that can operate in a 384-well format. The method is based on the inhibition in solution assay (ISA) approach using an immobilized target definition compound (TDC) that has been specifically designed for this purpose by making use of available structural information. We demonstrate that this method is robust and is sufficiently sensitive to detect fragment hits as weak as KD 500 μM when confirmed in a conventional surface plasmon resonance approach. The application of the plate-based screen, the identification of fragment inhibitors of PDE10A, and their structural characterization are all discussed in a forthcoming paper.

Full text of DOI:10.1021/jm301665y

Article
pubs.acs.org/jmc
Development of a Plate-Based Optical Biosensor Fragment
Screening Methodology to Identify Phosphodiesterase 10A
Inhibitors
Stefan Geschwindner,*,† Niek Dekker,† Rob Horsefield,† Anna Tigerstrom,† Patrik Johansson,†
̈
Clay W. Scott,‡ and Jeffrey S. Albert*,‡,§
†Discovery Sciences, AstraZeneca R&D Molndal, S-43183 Molndal, Sweden
̈
̈
‡CNS Discovery Research, AstraZeneca Pharmaceuticals, 1800 Concord Pike, PO Box 15437, Wilmington, Delaware 19850, United
States
S
* Supporting Information
ABSTRACT: We describe the development of a novel fragment screening methodology
employing a plate-based optical biosensor system that can operate in a 384-well format.
The method is based on the “inhibition in solution assay” (ISA) approach using an
immobilized target definition compound (TDC) that has been specifically designed for this
purpose by making use of available structural information. We demonstrate that this
method is robust and is sufficiently sensitive to detect fragment hits as weak as KD 500 μM
when confirmed in a conventional surface plasmon resonance approach. The application of
the plate-based screen, the identification of fragment inhibitors of PDE10A, and their
structural characterization are all discussed in a forthcoming paper.
of test compounds with the biosensor chip (e.g., super-
stoichiometric binding), limited assay sensitivity, and differ-
ences in dynamic range. Challenges often increase with (1)
increasing molecular weight of the target protein, (2)
decreasing molecular weight of the test compounds, and (3)
more weakly binding potential hits; all of these contribute to
limitations for the reliable detection of genuine binding events.9
Inhibition in Solution Assay (ISA) Screening Using
Surface Plasmon Resonance (SPR) with Conventional
Flow-Based Methods. For the optical biosensor approach,
various methodologies have been employed to address these
challenges discussed above.8 Typically, the macromolecular
protein target is immobilized on the biosensor. We have
recently employed an alternative strategy where a smaller
molecule probe compound is instead immobilized on the chip.
We refer to this tethered probe as the target definition
compound or TDC. The TDC must be immobilized to the
biosensor without compromising the binding to the target
protein. As illustrated in Figure 1, the TDC is immobilized on
the sensor surface. Next, a solution containing the enzyme
target is flowed across the sensor surface and that enzyme will
recognize and bind to the TDC causing a measurable signal.
That protein/TDC interaction can be modulated or disrupted
by coincubation of the enzyme with an inhibitor that is in
competition with the TDC for binding with the protein. In this
INTRODUCTION
■
Evolving Applications (and Challenges) in Fragment
Screening. Fragment-based drug design (FBDD) has become
widely used as a tool for discovering new chemical leads.1,2 In
comparison to conventional high throughput screening, the
alternative screening of low molecular weight, low complexity
compounds (fragments) allows more efficient sampling of
structure space3 and can be effective using screening libraries as
small as a few thousand compounds or less.4 This area has been
reviewed extensively.5−7 Fragment screening requires highly
sensitive, high quality assays in order to detect the typically
weak binding fragments. Efficient and effective fragment
screening activities are key to success in FBDD, but despite
the increasing importance and applications of FBDD, currently
employed screening approaches still suffer from distinct
limitations including complex assays, lower throughput, and
high false positive rates. A review of various methods and their
advantages and disadvantages has recently appeared.8
Various one-dimensional and two-dimensional NMR meth-
odologies have been developed and have been very successful
in fragment screening.9−11 However, NMR methods often have
low throughput and high reagent consumption. More recently,
optical biosensors and, in particular, surface plasmon resonance
(SPR) technologies, are increasingly being used as alternatives
to NMR screening.12−14 These methods offer increased
throughput (>1000 fragments/day) and lower reagent
consumption (<100 μg target protein). However, the SPR
approach has some challenges including unspecific interactions
Received: November 11, 2012
Published: March 19, 2013
© 2013 American Chemical Society
3228
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
concentration in the presence of a fragment, it enables the
detection of fragment binding in the millimolar affinity regime
and thus offers an attractive option for primary fragment
screening.
Developing a New Inhibition in Solution Assay (ISA)
Screening Approach Using Plate-Based OWG. The
conventional flow-based methods described above have limited
throughput due to the sequential application of samples onto
the biosensor. The advantage of the ISA approach when
employing high-throughput plate-based optical biosensor
platforms has been described already for the study of inhibitor
binding to human trypsin.16 Here we describe a modification of
the ISA approach adapted to a plate-based optical biosensor
methodology to provide a robust and sensitive assay to screen
fragments as binders/inhibitors of PDE10A. A forthcoming
report will describe the identification, characterization, and
structural elucidation of those fragment hits.
Inhibition in Solution Assay (ISA) Screening Method-
ology Using Optical Waveguide Grating (OWG). A range
of alternative and complementing biosensor platforms have
lately been developed to enable the study of protein−ligand
interactions with even higher throughput. Evanescent field
sensing provided by optical waveguides (also referred to as
optical waveguide grating or OWG) has emerged as a viable
complement to SPR platforms and have recently been
commercialized.17,18 Similar advantages and disadvantages as
compared to those already mentioned for SPR are also
applicable to these plate-based biosensor platforms; in
particular, they share the sensitivity limits associated with
working with immobilized target protein. Therefore, we sought
to adapt in ISA methodology to be used with an OWG system.
Such a system should provide the advantages of conventional
ISA/SPR technology but with substantially increased speed.
This approach is illustrated in Figure 2.
Figure 1. Schematic illustration of the ISA format used in SPR studies.
(A) The ISA format employs an immobilized TDC that ideally
displays high affinity to the enzyme. Competing inhibitors that are
present in solution will subsequently lower the observable binding
signal. (B) Illustration of the TDC (as employed in the present study)
linked to the sensor surface.
RESULTS AND DISCUSSION
■
manner, the inhibitor/protein interaction can be assessed
qualitatively as well as quantitatively.
TDC Development. In prior work, our colleagues identified
a series of high affinity PDE10A inhibitors.19,20 The binding
In general, good choices for compounds as TDCs are
substrate analogues or known inhibitors. A major criteria for
the design of a selective TDC is the retention of the biological
activity of the parent ligand including high affinity to the target
protein. The strategies for designing a suitable TDC can be
very much adopted from approaches aiming to develop
fluorescently labeled ligands to study ligand binding events.
This is a suitable approach in cases where detailed structural
information is not readily available, and one typically relies on
pharmacophore information and/or structure−activity relation-
ships for substrate analogues and/or known inhibitors for TDC
design. Nevertheless, this usually requires the synthesis of a
larger number of compounds that need to be tested in a trial
and error fashion in order to identify suitable candidates. The
potential availability of structural information furnishes the
opportunity to design such TDCs much more effectively and in
a rational fashion, thereby minimizing the risk of creating
unsuitable reagents during assay development.
Figure 2. Schematic illustration of the ISA format used in the present
study applying a plate-based OWG system. (1) In preparation for
screening, a TDC is covalently attached to the functionalized
biosensor surface. The incident light is reflected at a characteristic
wavelength. (2) Upon addition of the protein and its subsequent
interaction with the TDC, the resulting mass-change at the sensor
surface is leading to a wavelength increase of the reflected light
(typically 0.5−2 nm dependent on protein concentration). (3) The
addition of a competitive compound results in a negative mass-change
due to protein dissociation from the sensor surface and is manifested
in wavelength decrease of the reflected light.
To differentiate this assay format from the more conven-
tional direct binding assays, we termed this fragment screening
protocol “inhibition in solution assay” (ISA).15 The advantages
of ISA include (1) substantially increased sensitivity, (2)
generic and rapid assay development, and (3) immediate
verification of competitive binding. Because this method
measures precisely even small changes in the free protein
3229
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
mode was analyzed by X-ray crystallography for selected
representatives such as 1. From this structural information it
was evident that the position occupied by the hydroxyl pointed
out toward the solvent and could serve as the region to attach a
linker that could then be used to attach to the biosensor chip.
Next, we employed 2, which is a member of a related
chemotype21 that also had high affinity and a carboxylic acid as
an easily functionizable synthetic handle. The amide analogue 3
lost about 250-fold activity relative to 2; however, it still
retained sufficient potency to indicate that analogues containing
more extended linkers could be useful as tethered probes.
Molecular modeling of 3 (based on the structural information
for 1) indicated that a linker with length of about eight atoms
would be suitable to attach to the biosensor chip and still allow
subsequent binding of PDE10A to the probe. Accordingly, we
prepared 4, which had IC50 991 nM in our enzyme inhibition
assay and was thus judged to be suitable as the TDC. The
reduction in affinity is actually advantageous in order to enable
low-affinity binding fragments to effectively compete for the
same binding site. This comes at the cost of increasing protein
concentration in the assay in order to obtain a good signal/
noise-ratio. Thus it is important to find an optimal balance
between those two parameters that still enables cost-effective
screening within an affinity window that still allows the
detection of low-affinity fragment binding. The attachment of 4
to the biosensor chip is shown schematically in Figure 1B.
Validation of 4 as a TDC Using Surface Plasmon
Resonance (SPR). Several steps were taken during the
development of a sensitive inhibition in solution assay (ISA)
for the detection of fragment binding to PDE10A employing an
OWG platform using the SRU BIND system. The suitability of
TDC 4 for a fragment screening assay was validated by
tethering it to the dextran-matrix of an SPR Biosensor. The
experiments (data not shown) indicated a satisfactory
immobilization level (typically 450−600 RU) of 4 onto the
biosensor. Injection of a PDE10A solution demonstrated that
the immobilized compound can interact specifically with
PDE10A, as reflected by a large binding signal. Injecting
different concentrations of PDE10A and making use of the
mass transport limitations that was intentionally designed
through the experimental setup (high immobilization density
and low flow rate), we could show that the observed binding
signal was solely determined by the mass transfer of the protein
to the sensor surface and resulted in a concentration-dependent
response. This setup could be used later on for the validation of
the initial fragment hits as well as for their affinity
determination in a similar fashion as reported earlier.8,15
Optical Waveguide Grating (OWG) Assay. In principle,
the above SPR methodology could have been used to screen
the fragment library. However, on the basis of previous
experience that some compounds/fragments can also bind
promiscuously to the immobilized TDC (thus impacting
negatively on subsequent ligand binding experiments), we
wanted to make use of the ability to measure a single
compound/fragment in a designated well of a 384-well plate in
a parallel fashion using a plate-based OWG platform. The
simultaneous investigation of 384 individual interactions is a
vast increase in throughput in comparison to iterative
measurements using flow-based SPR systems. Furthermore,
we wanted to conduct the screening as efficiently as possible
without exposing the protein to the fragments during extended
time periods to prevent time-dependent precipitation due to
solubility problems or promiscuous binding.
For the development of the ISA on the OWG platform, we
found it necessary to optimize two key parameters in order to
define an optimal dynamic window for the detection of
fragment binding: (1) the density of the TDC on the biosensor
surface and (2) the protein concentration. For this, we
evaluated a small matrix of different immobilization densities
as defined by the concentration of the ligand (1 mM to 50 μM)
during immobilization versus different protein concentrations
(10 μM to 100 nM); this was greatly facilitated by the plate-
based format of the OWG platform. This matrix was tested
against a range of 20 PDE10A inhibitors that are simpler
analogues of 5 (see Supporting Information) at a concentration
of 20 μM spanning different molecular weights from 200 to 400
Da and affinities from KD = 40 nM to 500 μM (as been
previously determined by employing the SPR ISA format). We
found that a high immobilization density of the TDC generally
led to increased unspecific binding of the tested inhibitors to
the TDC and that the displacement effects became less
pronounced for inhibitors with weak affinity, irrespective of the
protein concentration. As expected, the protein concentration
plays an essential role in defining the dynamic window of the
assay, as the readable signal usually gradually increases with
increasing protein concentration. However, in analogy to the
density of the TDC, the displacement effects at relatively high
protein concentrations of 5−10 μM became less pronounced in
conjunction with a diminished increase of the protein binding
signal. To address this issue, we chose not to work with the
maximal attainable protein binding signal but rather to reduce
the protein concentration used in the assay as well as the
concentration of the TDC during immobilization. By applying
this matrix, we found an optimal combination of TDC density
(100 μM immobilization concentration for 30 min leading to
approximately 1500 ΔPWV immobilization signal) and protein
3230
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
concentration (2 μM) that enabled the detection of fragment
binding with very low affinities (close to millimolar) using a
screening concentration of only 100 μM. Using these optimized
screening conditions, we compared the capacity of the already
mentioned PDE10A inhibitors to displace the protein from the
TDC-modified biosensor surface with their affinity as
determined via SPR ISA. As shown in Figure 3B, we could
confirm a good correlation between the reduction in PDE10A
binding to the biosensor with increasing KD values, indicating
the possibility to conduct an reliable affinity ranking that is
solely based on the reduction in PDE10A binding to the TDC-
modified biosensor. To assess the binding specificity in this
setup, we employed 1 using a concentration that leads to full
occupancy of the binding site (20 μM). We could observe
almost full displacement (>98%) of PDE10A from the
Biosensor surface (data not shown), indicating a highly specific
binding of PDE10A to the immobilized TDC.
Fragment Screening. To illustrate the experimental setup,
the time-course of the signals from several wells of a 384-well
plate are shown in Figure 3A. After the equilibration of all the
wells, a PDE10A solution is added and the specific interaction
of the TDC with PDE10A is reflected in a significant increase
of the signal which reaches a new equilibrium signal that is
arbitrarily set to 100%. The subsequent addition of compounds
that are competing with the PDE10A−TDC interaction
influences the equilibrium and is reflected in a significant
decrease of the binding signal corresponding to the affinity of
the compounds (Figure 3B).
CONCLUSIONS
■
In summary, we have described a new in-solution assay (ISA)
approach making use of plate-based optical biosensors that
offers an attractive option to SPR. The methodology is based
on derivatizing the sensor chip with a tool compound or TDC
that selectively recognizes and binds to the target protein at a
well-defined site. By making use of structural information, we
could drive rational design and synthetic modification of a tool
compound by optimal positioning of a suitable functional group
and linker to allow for the controlled immobilization of the
compound while retaining its high binding affinity to the target;
PDE10A. This methodology greatly simplifies assay develop-
ment and improves robustness. The displacement of a large
molecule from the biosensor surface in the presence of a
fragment offers an increased dynamic range that allows to
reliably detect millimolar binding events at concentrations that
are an order of magnitude lower. This in combination with the
high-throughput capability of OWG systems appears to be an
attractive alternative for primary fragment screening. The use of
this system to identify fragment hits for PDE10A, as well as
their further characterization and structural elucidation, will be
described in a forthcoming report.
Figure 3. Representative results from the OWG ISA. (A) Shown are
typical sensorgrams (example of 14 compounds displaying different
potencies at 20 μM and highlighted by different colors) illustrating the
experimental setup of the OWG ISA. After buffer equilibration, the
addition of PDE10A to TDC-containing wells at 7 min results in
specific binding to the biosensor surface. Addition of protein to
unmodified wells is used as control and to assess specificity. The
addition of compounds after equilibration at 33 min results in a
displacement of PDE10A from the biosensor surface due to specific
competition with the TDC. The magnitude of the displacement
reflects the potency of the compounds. (B) Correlation of compound
activity in the OWG ISA and compound affinity determined via SPR
ISA in the test set of 20 PDE10A inhibitors (details in Supporting
Information). In (A) and (B), the binding responses are shown as
PDE10A binding %, i.e., the binding of PDE10A to the TDC-modified
biosensor in the presence of fragment in relation to the controls
containing only PDE10A protein. PDE10A binding percentage is
defined by the OWG signal prior to addition of PDE10A (defined as
0% binding) in relation to the equilibrated OWG signal after addition
of PDE10A at the given concentration (defined as 100% binding). The
magnitude of the displacement correlates well with the affinity of the
compounds (linear regression line depicted in black) and can thus be
used for selection of compounds and preliminary affinity ranking.
EXPERIMENTAL SECTION
■
General Synthetic Methods. All reagents and solvents were
purchased from commercial sources and used without further
purification. Nonaqueous reactions were performed under a nitrogen
atmosphere. Reactions conducted in aqueous media were run under an
ambient atmosphere unless otherwise noted. Microwave irradiation
was carried out in a Personal Chemistry Emrys Optimizer microwave.
Flash column chromatography was performed using silica gel (MP,
32−63 μm, 60 Å) or using silica gel cartridges (GraceResolv or
RediSep normal phase disposable flash columns) on an ISCO
Companion. Reverse phase preparative HPLC purification was
performed with an Agilent 1200 LC using a Gemini C18 column
(250 mm × 21.2 mm, 10 μm particle diameter). Conditions: mobile
phase A was 2.5 mM aqueous ammonium bicarbonate solution, and
mobile phase B was acetonitrile. A gradient was run from 30 to 90% B
over 16 min at a column flow rate of 65 mL/min. UV detection was
recorded at 254 and 220 mm. All final compounds were determined to
be ≥95% purity by LCMS and 1H NMR unless specifically mentioned.
Proton NMR spectra were recorded on a Bruker Avance 500 MHz
spectrometer or a Bruker DRX 300 MHz spectrometer. Chemical
shifts are reported in ppm relative to tetramethylsilane as an internal
standard. Apparent peak multiplicities are described as s (singlet), br s
3231
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
(broad singlet), d (doublet), dd (doublet of doublets), t (triplet), q
(quartet), or m (multiplet). Coupling constants (J) are reported in
hertz (Hz). High resolution LCMS analyses were conducted on an
Agilent TOF 6210 MS using a Sorbax SB-C8 column (2.1 mm × 30
mm, 1.8 μm particle diameter) at 60 °C (Agilent 1200 LC).
Conditions: mobile phase A was water:acetonitrile:formic acid
(98:2:0.1 v/v/v), and mobile phase B was water:aceteonitrile:formic
acid (2:98:0.05 v/v/v). A linear gradient was run from 5 to 95% B over
1.5 min and held at 95% for 0.4 min. The column flow rate was 1.2
mL/min. Diode array UV detection was an averaged signal recorded
from 210 to 320 nm. Retention time (tR) was reported in minutes
(min). All final compounds that were tested in biological assays had
purity >95% as determined by analytical HPLC using at least two
different sets of conditions.
was concentrated under reduced pressure. The resulting aqueous
material was further concentrated with additional methanol (3×) to
provide a yellow oil that solidified when placed under vacuum. The
solid hydrochloride salt was taken up in methanol (5 mL), treated with
MP-carbonate (0.5 g, Aldrich, typical loading 2.5−3.5 mmol N/g)),
and stirred at room temperature for 30 min. The suspension was
filtered and then concentrated under reduced pressure to an oil. The
crude material was purified by flash column chromatography on silica
gel using a gradient mixture of 0−10% solvent A in solvent B, where
solvent A was a 2 M solution of ammonia in methanol and solvent B
was dichloromethane, to provide the product as an oil. The oil was
taken up in 1:1 acetonitrile/water (2 mL) and lyophilized to provide
1
the desired product as a solid (0.043 g, 89%). H NMR (500 MHz,
CDCl3) δ 1.37 (t, J = 7.0 Hz, 3 H), 1.52 (br s, 2 H), 2.50 (s, 3 H), 2.87
(br s, 2 H), 3.56 (t, J = 5.2 Hz, 2 H), 3.70−3.89 (m, 4 H), 4.08 (q, J =
7.0 Hz, 2 H), 6.66 (t, J = 5.3 Hz, 1 H), 6.94−7.10 (m, 2 H), 7.28−7.36
(m, 1 H), 7.38 (dd, J = 7.5, 1.7 Hz, 1 H), 7.41−7.53 (m, 2 H), 7.58 (d,
J = 8.2 Hz, 2 H), 7.69 (d, J = 8.2 Hz, 2 H), 8.13 (dd, J = 9.2, 5.2 Hz, 1
H). MS m/z (ES+), (M + H)+, HRMS (calcd) for C29H30FN3O3 =
488.23440; HRMS (obs) = 488.23447. HPLC tR = 1.37 min.
Cloning and Expression of PDE10A 449−789. The
pET21MDX7/hPDE10A clone was cotransformed with protein
chaperon vector pKJE8 (TaKaRa Bio Inc.) into BL21-Gold(DE3)
cells. Coexpression was done following the protocols provided by
TaKaRa Bio Inc. Expression of protein chaperons was induced using 1
mg/mL arabinose and 5 ng/mL tetracycline at a culture OD600 of 0.3.
Expression of target protein was induced by an addition of 0.1 mM
IPTG at an OD600 of 0.6. After induction, the cell culture was grown
at 16 °C overnight.
2-(2′-Ethoxybiphenyl-4-yl)-6-fluoro-3-methylquinoline-4-carbox-
ylic Acid1 (2). 2 was prepared from 5-fluoroindoline-2,3-dione and 1-
(4-bromophenyl)propan-1-one using the Pfitzinger reaction2 accord-
ing to literature methods. Compound 2 was coupled with N-Boc-2,2′-
oxybis(ethylamine) (3a) and then deprotected to afford 4 (Scheme 1).
a
Scheme 1
Protein Purification. Cells were harvested and resuspended into
lysis buffer (50 mM Tris, pH 7.5, 300 mM NaCl, 1 mM DTT) at the
ratio of 5 mL of lysis buffer per gram of cells. Then 1 mM MgCl2 and 1
mM PMSF was added to the suspension before sonication on ice (5 s
on-time, 10 s off-time, total 3 min on-time). Cell lysate was centrifuged
at 13000 rpm for 60 min at 4 °C. The supernatant was loaded onto a
pre-equilibrated 5 mL Ni-NTA column and washed with 10 CV of
lysis buffer followed by 20 CV of lysis buffer with 20 mM imidazole.
The target protein was subsequently eluted with lysis buffer containing
300 mM imidazole. Protein fractions were collected and analyzed by
SDS-PAGE. Fractions containing PDE10A protein were diluted six
times using a 50 mM Tris-Cl (pH 7.5) buffer. ZnCl2, MgCl2, AMP,
and β-mercaptoenthanol were added to final concentrations of 0.1
mM, 0.1 mM, 1 mM, and 10 mM, respectively. The protein sample
was loaded onto a 5 mL HiTrap-QP column (GE Healthcare Life
Sciences) pre-equilibrated with 50 mM Tris-Cl, pH 7.5, 50 mM NaCl,
and 10 mM β-mercaptoenthanol. A 0−50% gradient of 50 mM Tris-
Cl, pH 7.5, 2 M NaCl, 10 mM β-mercaptoenthanol was used for
elution. The flow-through was collected for further purification.
Fractions containing the target protein were digested with TEV
protease at 4 °C overnight while dialyzing against the lysis buffer.
Completeness of His-tag cleavage was assessed by SDS-PAGE. As a
last purification step, the protein sample was loaded to a pre-
equilibrated Ni-NTA column. The flow-through from loading and
washing with lysis buffer containing 5, 10, and 20 mM imidazole was
collected. After analysis by SDS-PAGE, the purest fractions were
pooled.
SPR Inhibition in Solution Assay for Fragment Validation
and KD Estimation. A tethered free amine ligand 3 was covalently
linked to the dextran layer on the CM5 sensor chip (GE/Biacore)
using the Amine Coupling Kit (GE/Biacore) and performed according
to the manufacturer’s recommendations. Carboxyl groups on the
dextran layer were activated by injecting a 1:1 mixture of 0.5 M N-
ethyl-N′-(3-dimethylaminopropyl)-carbodiimide (EDC) and 0.5 M N-
hydroxysuccinimide (NHS) at 5 μL/min for 7 min. The activated
carboxyl groups on the control channels were blocked by injecting a 1
M ethanolamine solution for 7 min at 5 μL/min. The tethered ligand
was coupled to the biosensor surface by injecting a 100 μM solution of
the compound (HBS-P buffer plus 2% DMSO) over the chip surface
at 10 μL/min for 12 min. The remaining activated groups were
blocked with 1 M ethanolamine. Compounds were tested at either a
a
Reagents and conditions: (a) TBTU, HOBT, DIPEA, DMF, RT,
66%; (b) HCl, 89%.
tert-Butyl 2-(2-(2-(2′-Ethoxybiphenyl-4-yl)-6-fluoro-3-methylqui-
noline-4-carboxamido)ethoxy)ethylcarbamate (2). A mixture of 2-
(2′-ethoxybiphenyl-4-yl)-6-fluoro-3-methylquinoline-4-carboxylic acid1
(1) (0.075 g, 0.19 mmol), tert-butyl 2-(2-aminoethoxy)-
ethylcarbamate3 (0.064 g, 0.31 mmol), and HOBT hydrate (0.036 g,
0.24 mmol) in DMF (2.5 mL) was treated with TBTU (0.076 g, 0.24
mmol) and diisopropylethylamine (0.100 mL, 0.57 mmol) and then
stirred at room temperature overnight. Additional TBTU (0.035 g)
was added, and stirring continued for 2 h. DMF was removed under
vacuum. The concentrate was partitioned between aqueous potassium
carbonate solution and ethyl acetate. The organic portion was washed
(water, brine), dried (magnesium sulfate), filtered, and evaporated.
The residue was purified by flash column chromatography on silica gel
using 0−50% ethyl acetate in hexane as eluent to provide the desired
1
product as an oil (0.073 g, 66%). H NMR (500 MHz, CDCl3) δ
1.20−1.35 (m, 9 H), 1.37 (t, J = 7.0 Hz, 3 H), 2.51 (s, 3 H), 3.27−3.36
(m, 2 H), 3.58 (t, J = 5.2 Hz, 2 H), 3.71−3.79 (m, 2 H), 3.78−3.86
(m, 2 H), 4.08 (q, J = 7.0 Hz, 2 H), 4.75 (br s, 1 H), 6.58 (br s, 1 H),
6.97−7.09 (m, 2 H), 7.27−7.35 (m, 1 H), 7.38 (dd, J = 7.5, 1.7 Hz, 1
H), 7.41−7.52 (m, 2 H), 7.58 (d, J = 8.2 Hz, 2 H), 7.69 (d, J = 8.2 Hz,
2 H), 8.13 (dd, J = 9.0, 5.3 Hz, 1 H). MS m/z (ES+), (M + H)+,
HRMS (calcd) for C34H38FN3O5 = 588.28683; HRMS (obsd) =
588.28583. HPLC tR = 1.68 min.
N-(2-(2-Aminoethoxy)ethyl)-2-(2′-ethoxybiphenyl-4-yl)-6-fluoro-
3-methylquinoline-4-carboxamide (3). A solution of tert-butyl 2-(2-
(2-(2′-ethoxybiphenyl-4-yl)-6-fluoro-3-methylquinoline-4-
carboxamido)ethoxy)ethylcarbamate (2) (0.058 g, 0.10 mmol) in
methanol (3.0 mL) was treated with 3N aqueous hydrochloric acid
(3.0 mL, 9.0 mmol) and concentrated hydrochloric acid (0.5 mL, 6.0
mmol). After stirring at room temperature for 1 h, the reaction mixture
3232
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
single concentration or in a concentration response curve. For the
concentration response curve, compounds were serially diluted 1:4 in
DMSO to form a seven-point concentration response curve. To a 96-
well plate, 1.4 μL/well compound solutions were spotted. The
PDE10A catalytic domain at 170 nM was added to the plate and the
mixture incubated for 15 min at RT, and then the plate was sealed and
placed in the Biacore 3000 instrument. Samples were injected at 35
μL/min within 17 s. The signal generated on the control channel was
subtracted from the signal of the channel with the tethered ligand. The
value used to calculate percent inhibition of binding was the slope of
the response measured 8 s after the start of the injection. The surface
of the chip was regenerated by injecting 0.5% SDS for a period of 15 s.
OWG Inhibition in Solution Assay for Fragment Screening.
The tethered free amine ligand 3 was covalently linked to the activated
surface of a GA4 384-well biosensor plate (SRU Biosystems). Prior to
activation, the biosensor plate was washed with deionized water. A
mixture of N-ethyl-N′-(3-dimethylaminopropyl)-carbodiimide (EDC)
and N-hydroxysuccinimide (NHS) was prepared by mixing 1.9 mL of
a 10 mM NHS-solution with 0.2 mL of a 650 mg/mL solution of EDC
and adding 17.9 mL of deionized water. From this mixture, 20 μL was
added into each well of the 384-well biosensor plate and incubated for
15 min at RT. After incubation, all liquid was quickly removed to
apparent dryness by flick and tap and the tethered ligand was coupled
to the biosensor surface by adding 20 μL of a 100 μM ligand solution
in 10 mM Na-acetate, pH 5.60, to the individual wells. After 30 min
incubation, the biosensor plate was washed with HBS-P buffer to
remove excess ligand and 20 μL of 0.4 M ethanolamine was added to
block remaining activated groups for 30 min. Thereafter, the plate was
washed extensively with HBS-P buffer and the biosensor was finally
stabilized with HBS-P plus 1% DMSO. Some wells of the plate have
been selected as control wells by omitting the addition of the ligand
solution after the activation step.
The fragment screening was performed by first preparing a 1 mM
solution of the selected fragments in HBS-P plus 1% DMSO. The
biosensor plate was equilibrated with 22.5 μL of HBS-P plus 1%
DMSO, and a baseline read was obtained with a sampling rate of two
data points per minute. Then 4.5 μL of a 12 μM PDE10A solution in a
matching buffer was added, and a new equilibrium signal was obtained
after 25−30 min, which was typically within the range of 500−600 pm.
Thereafter, 3 μL of the fragment solution was added to reach a final
fragment concentration of 100 μM. A new equilibrium signal was
obtained after another 5−10 min. The displacement of PDE10A from
the tethered compound upon addition of the fragment was reported as
a percent of residual binding of PDE10A in relation to the maximum
equilibrium signal. For the concentration response curve, increasing
amounts of compounds were sequentially added into each individual
well of the plate and the new equilibrium signal after each addition was
used to calculate the percent of residual binding in order to form a
four-point concentration response curve. This was used to calculate
the IC50 values.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS USED
■
FBDD, fragment-based drug design; ISA, inhibition in solution
assay; NMR, nuclear magnetic resonance; OWG, optical
waveguide grating; PDE10A, phosphodiesterase 10A; PWV,
peak wavelength value; SPR, surface plasmon resonance; TDC,
target definition compound
REFERENCES
■
(1) Hajduk, P. J.; Greer, J. A decade of fragment-based drug design:
strategic advances and lessons learned. Nature Rev. Drug Discovery
2007, 6, 211−219.
(2) Congreve, M.; Chessari, G.; Tisi, D.; Woodhead, A. J. Recent
developments in fragment-based drug discovery. J. Med. Chem. 2008,
51, 3661−3680.
(3) Hann, M. M.; Leach, A. R.; Harper, G. Molecular complexity and
its impact on the probability of finding leads for drug discovery. J.
Chem. Inf. Comput. Sci. 2001, 41, 856−864.
(4) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful
metric for lead selection. Drug Discovery Today 2004, 9, 430−431.
(5) Hubbard, R. E.; Murray, J. B. Experiences in fragment-based lead
discovery. Methods Enzymol. 2011, 493, 509−531.
(6) Leach, A. R.; Hann, M. M. Molecular complexity and fragment-
based drug discovery: ten years on. Curr. Opin. Chem. Biol. 2011, 15,
489−496.
(7) Lepre, C. A. Practical aspects of NMR-based fragment screening.
Methods Enzymol. 2011, 493, 219−239.
(8) Holdgate, G. A.; Anderson, M.; Edfeldt, F.; Geschwindner, S.
Affinity-based, biophysical methods to detect and analyze ligand
binding to recombinant proteins: matching high information content
with high throughput. J. Struct. Biol. 2010, 172, 142−157.
(9) Dalvit, C. NMR methods in fragment screening: theory and a
comparison with other biophysical techniques. Drug Discovery Today
2009, 14, 1051−1057.
(10) Hajduk, P. J.; Olejniczak, E. T.; Fesik, S. W. One-Dimensional
Relaxation- and Diffusion-Edited NMR Methods for Screening
Compounds That Bind to Macromolecules. J. Am. Chem. Soc. 1997,
119, 12257−12261.
(11) Jahnke, W.; Erlanson, D. A. Fragment-Based Aproaches to Lead
Discovery; Wiley: Weinheim, Germany, 2007.
(12) Danielson, U. H. Fragment Library Screening and Lead
Characterization using SPR Biosensors. Curr. Top. Med. Chem. 2009,
9, 1725−1735.
(13) Giannetti, A. M. From experimental design to validated hits: a
comprehensive walk-through of fragment lead identification using
surface plasmon resonance. Methods Enzymol. 2011, 493, 169−218.
(14) Geschwindner, S. Industrial biophysics at the outlook for better
leads: accelerating lead generation using affinity-based, label-free
methods. Int. Drug Discovery 2010, 5 (32), 34−36.
(15) Geschwindner, S.; Olsson, L.-L.; Albert, J. S.; Deinum, J.;
Edwards, P. D.; de Beer, T.; Folmer, R. H. A. Discovery of a Novel
Warhead against beta-Secretase through Fragment-Based Lead
Generation. J. Med. Chem. 2007, 50, 5903−5911.
(16) Geschwindner, S.; Carlsson, J. F.; Knecht, W. Application of
Optical Biosensors in Small-Molecule Screening Activities. Sensors
2012, 12, 4311−4323.
(17) Cunningham, B. T.; Li, P.; Schulz, S.; Lin, B.; Baird, C.;
Gerstenmaier, J.; Genick, C.; Wang, F.; Fine, E.; Laing, L. Label-free
assays on the BIND system. J. Biomol. Screening 2004, 9, 481−490.
(18) Fang, Y.; Ferrie, A. M.; Fontaine, N. H.; Mauro, J.; Balakrishnan,
J. Resonant waveguide grating biosensor for living cell sensing. Biophys.
J. 2006, 91, 1925−1940.
(19) Bauer, U.; Giordanetto, F.; Bauer, M.; O’Mahony, G.; Johansson
Kjell, E.; Knecht, W.; Hartleib-Geschwindner, J.; Carlsson Eva, T.;
Enroth, C. Discovery of 4-hydroxy-1,6-naphthyridine-3-carbonitrile
ASSOCIATED CONTENT
■
S
* Supporting Information
Information and assay data for the used inhibitors for assay
validation. Result of the molecular modeling of 3 based on the
structural information for 1. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*For S.G.: phone, +46 31 776 2197; fax, +46 31 776 3792; E-
mail, Stefan.Geschwindner@astrazeneca.com. For J.A: E-mail,
Jeffrey.Albert@IntelliSynRD.com.
Present Address
§IntelliSyn R&D, 201 President Kennedy Avenue, Montreal,
QC H2X 2J6, Canada.
3233
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234

Journal of Medicinal Chemistry
Article
derivatives as novel PDE10A inhibitors. Bioorg. Med. Chem. Lett. 2012,
22, 1944−1948.
(20) O’Mahony, G.; Bauer, U. Unpublished results.
(21) Cantin, D.; Magnuson, S.; Gunn, D.; Bullock, W.; Burke, J.; Fu,
W.; Kumarasinghe, E. S.; Liang, S. X.; Newcom, J.; Ogutu, H.; Olague,
A.; Wang, M.; Wickens, P.; Zhang, Z.; Bierer, D. Preparation of
phenyl-substituted quinoline and quinazoline amino acid derivatives
for the treatment of diabetes. Patent WO2006034512, 2006.
3234
dx.doi.org/10.1021/jm301665y | J. Med. Chem. 2013, 56, 3228−3234