Atropisomerism and Conformational Equilibria: Impact on PI3Kδ Inhibition of 2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-(o-tolyl)quinazolin-4(3H)-one (IC87114) and Its Conformationally Restricted Analogs

Article abstract of DOI:10.1021/acs.jmedchem.7b00247

IC87114 [compound 1, (2-((6-amino-9H-purin-9-yl)methyl)-5-methyl-3-(o-tolyl)quinazolin-4(3H)-one)] is a potent PI3K inhibitor selective for the δ isoform. As predicted by molecular modeling calculations, rotation around the bond connecting the quinazolin-

Full text of DOI:10.1021/acs.jmedchem.7b00247

Article
pubs.acs.org/jmc
Atropisomerism and Conformational Equilibria: Impact on PI3Kδ
Inhibition of 2‑((6-Amino‑9H‑purin-9-yl)methyl)-5-methyl-
3‑(o‑tolyl)quinazolin-4(3H)‑one (IC87114) and Its Conformationally
Restricted Analogs
Alessio Lodola,† Serena Bertolini,‡ Matteo Biagetti,‡ Silvia Capacchi,‡ Fabrizio Facchinetti,‡
Paola Maria Gallo,‡ Alice Pappani,‡ Marco Mor,*,† Daniele Pala,†,∥ Silvia Rivara,† Filippo Visentini,§
Mauro Corsi,§ and Anna Maria Capelli*,‡
†Dipartimento di Scienze degli Alimenti e del Farmaco, Universita
̀
degli Studi di Parma, Viale delle Scienze 27/A, 43124 Parma, Italy
‡Chemistry Research and Drug Design Department, Chiesi Farmaceutici S.p.A., Largo F. Belloli 11/A, 43122 Parma, Italy
§Aptuit s.r.l., Via Fleming 4, 37135 Verona, Italy
S
* Supporting Information
ABSTRACT: IC87114 [compound 1, (2-((6-amino-9H-purin-9-yl)methyl)-5-methyl-3-(o-tolyl)quinazolin-4(3H)-one)] is a
potent PI3K inhibitor selective for the δ isoform. As predicted by molecular modeling calculations, rotation around the bond
connecting the quinazolin-4(3H)-one nucleus to the o-tolyl is sterically hampered, which leads to separable conformers with axial
chirality (i.e., atropisomers). After verifying that the aS and aR isomers of compound 1 do not interconvert in solution, we
investigated how biological activity is influenced by axial chirality and conformational equilibrium. The aS and aR atropisomers of
1 were equally active in the PI3Kδ assay. Conversely, the introduction of a methyl group at the methylene hinge connecting the
6-amino-9H-purin-9-yl pendant to the quinazolin-4(3H)-one nucleus of both aS and aR isomers of 1 had a critical effect on the
inhibitory activity, indicating that modulation of the conformational space accessible for the two bonds departing from the central
methylene considerably affects the binding of compound 1 analogues to PI3Kδ enzyme.
The decision to develop a drug candidate as an atropisomer
should be in place as early as possible during the lead
optimization stage as the existence of multiple atropisomers
would even further complicate their drug development in terms
of stability on the shelf and potential risk of interconversion
after administration. Despite these critical issues, no direct
guidelines as to how deal with time-dependent chirality have
been delivered by regulatory agencies.8 On the other hand,
pharmaceutical companies have developed theoretical and
experimental protocols to characterize atropisomer intercon-
version in order to facilitate well-versed decisions about storage
and handling of compound under pharmaceutical develop-
ment.9,10
INTRODUCTION
■
Atropisomerism is a source of molecular asymmetry arising
from the presence of a restricted rotation along a covalent bond
in a given structure, giving rise to stable conformers with axial
chirality (atropisomers) that can be isolated as distinct chemical
species when their interconversion half-life is sufficiently long,
e.g., higher than 1000 s.1 Drugs containing classical stereogenic
centers are considered stable compounds as they can only
racemize through bond breaking processes. Conversely,
atropisomers racemize through a simple bond rotation with a
time scale ranging from minutes to years, depending on steric
hindrance, electronic factors, temperature, and solvent.
Atropisomeric compounds, like classical enantiomers, are
featured by different pharmacodynamics and pharmacokinetics
profiles.2,3 Indeed, for several atropisomer pairs, significant
variation in in vitro inhibition or binding data4 as well as in
ADMET properties5 have been reported in the literature.6,7
Received: February 14, 2017
Published: May 10, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Chart 1. Chemical Structures of PI3Kδ Inhibitors Considered in This Study
We recently focused our attention on 2-((6-amino-9H-purin-
9-yl)methyl)-5-methyl-3-(o-tolyl)quinazolin-4(3H)-one,
(IC87114, compound 1, Chart 1), a potent and selective
inhibitor of the phosphoinositide 3-kinase (PI3K) δ isoform,11
an enzyme that has been implicated in several pathological
conditions, including autoimmune and inflammatory diseases
and B-cell malignancies. Since compound 1 has been widely
utilized as tool compound to carry out both in vitro and in vivo
experimental studies on PI3Kδ, its in-depth characterization
may help to better interpret the data generated so far.
For this derivative, the rotation along the bond connecting
the quinazolin-4-one moiety to the N-tolyl fragment appears to
be hindered, suggesting that 1 may exist as a mixture of two
separable atropisomers, i.e., the aS isomer, where the methyl
group of the tolyl group lies above the plane of the quinalozin-
4-one ring (Chart 1, compound 2), and the aR isomer (Chart 1,
compound 3), where the methyl group lies on the opposite side
of that plane.
atropisomers. Our work started with the characterization of a
commercial batch of 1, isolation of atropoenantiomers, and
their identification by vibrational circular dicroism (VCD). The
extent of mutual conversion of 2 and 3 was measured at
conventional temperatures to assess the risk that, starting from
an atropoenantiomer, racemization could occur during
compound manipulation and biological tests.
Atropisomerism is a particular feature of conformational
space, where energetic barriers are too high to be freely crossed
at room temperature. While equilibria among conformations
separated by low energy barriers can be investigated by free-
energy calculations employing molecular mechanics force fields,
higher barriers, like those involved in atropisomerism, require
contributions from quantum mechanics (QM).8,9 Estimation of
the barrier for rotation about the bond involved in axial chirality
was qualitatively compared to stability data to confirm the
predictive ability of QM-based computational protocols.
In a more general sense, equilibria among accessible
conformations of a drug can affect its propensity to interact
with specific targets, and the ability to predict the effect of
chemical modifications on biological properties is essential to
drug design and development. Applications of computational
protocols to the study of the conformational space of drugs and
investigations on its relevance for drug−protein interactions
have been reported in the literature.15,16Available information
about the inhibitor mechanism of 1 on PI3Kδ suggests that
reciprocal arrangements of the tolylquinazolinone and adenyl
moieties are essential for binding at the active site.17 Thus, we
combined computational and experimental techniques to
explore in a wider sense the relationship between the
conformational space of compound 1, with its close derivatives,
and affinity for PI3Kδ. The conformational free-energy surface
(FES) of a bioactive compound provides critical information
for drug design purposes. Knowledge of the penalty (i.e., strain-
energy, ΔEstrain) that a ligand has to pay for leaving the global-
minimum basin in solution and adopting the target-binding
conformation18 is instrumental to design conformationally
In fact, a patent application12 describes a method to obtain,
by chiral chromatography, the two optically active atropoe-
nantiomers 2 and 3, reporting similar IC50 values for PI3Kδ
inhibition and different pharmacokinetics. In the same patent,
the first eluting compound was identified as the aS atropisomer
2 by X-ray diffraction. Prompted by this report, we decided to
explore the composition of the commercial batches of 1, predict
and measure the stability of each atropisomer, and investigate
the influence of steric effects related to the conformational
space of 1 on PI3Kδ inhibitor potency.
Although compound 1 is regarded as one of the reference
inhibitors of PI3Kδ enzyme, for which several bioanalytical
methods have been recently validated to determine its
plasmatic level in mice,13,14 no data about its axial chirality
are provided for commercial batches. As atropisomerism often
goes both undetected and unsuspected given that enantiomers
cannot be distinguished using routine analytical tools (1H
NMR, HPLC, etc.),8,9 it might be the case that the compound
employed for biological tests is actually a racemic mixture of
B
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
restricted compounds exhibiting smaller ΔEstrain, hence
expected to be more potent than their flexible analogues.
The X-ray crystal structure of a rotamer, corresponding to
the putative aS atropisomer of 1 (i.e., compound 2), bound to
the catalytic domain of human PI3Kδ (PDB code 2X38)17
shows that the adenine moiety of this ligand takes hydrogen-
bond interactions with the hinge amino acids, whereas the
quinazoline fragment is accommodated in the “specificity
pocket” beneath the enzyme P-loop. The tolyl group is located
at the entrance of the cavity where the binding pocket is wide
and solvent exposed (Figure 1).17
Visual inspection of this complex and molecular docking (not
shown) suggest that the rotamer with opposite orientation of
the tolyl fragment (corresponding to the putative aR
atropisomer, structure 3) can also be accommodated into the
ATP-binding site of PI3Kδ. While the absence of atropo-
stereoselectivity can be anticipated by similar potency data
reported for the two atropisomers in the cited patent, we
speculated that modulation of the conformational space
accessible for the two bonds departing from the central
methylene could affect binding site activity. In particular, the
introduction of a methyl group on the methylene spacer
connecting the adenine ring to the quinazolinone nucleus of 2
and 3 (leading to the atropodiastereoisomer couples 5−6 and
7−8) could restrict the conformational space of these
compounds compared to that of the reference inhibitor 1 and
thus affect their inhibitory potency. Hence, an investigation of
the conformational space of compound 2 and its analogues
together with the assessment of their biological profile appears
fundamental to evaluate the impact of atropisomerism and
conformational equilibria on PI3Kδ inhibition.
Here we report the characterization of compound 1 as an
atropisomeric mixture, the isolation of pure aS and aR
atropisomers followed by their extensive characterization in
terms of stability, conformational preferences, and biological
profile. Finally, a few conformationally restricted analogues
characterized by the presence of both axial chirality and
stereogenic centers were in depth investigated.
RESULTS AND DISCUSSION
■
Chemistry. Compound 1 was purchased from Symansis,
while single atropisomers (compound 2 and compound 3)
were obtained from HPLC separation of the racemate (see
Experimental Section for details). The preparation of
Figure 1. aS atropisomer of 1 (compound 2, green carbon atoms) in
complex with PI3Kδ (white carbon atoms) as revealed by X-ray
crystallography (PDB code 2X38).17
a
Chart 2. Synthesis of Compound 4 and the Racemic 13
a
Reagents and conditions: (i) SOCl2 (3.0 equiv), toluene, 18 h, reflux; (ii) R1NH2 (1.5 equiv), TEA (1.5 equiv), DCM, 2 h, reflux; (iii) Cl-
CHR2COCl (11a, R1 = H; 11b, R1 = Me), AcOH, reflux, 18 h ; (iv) adenine (1.5 equiv), K2CO3 (1.5 equiv), acetonitrile.
C
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. Left: Splitting of CH2, signals observed in the 1H NMR spectra of the association complex 1:R-Pirkle. Right: Splitting of signals observed in
the 1H NMR spectra of the association complex 1:R-Pirkle for the two methyl groups on the quinazolinone nucleus (2.8 ppm) and on the phenyl
ring (2.1 ppm). Letters a and b indicate the two atropisomers.
quinazolinone analogues of 1, namely, compound 4 and the
racemic mixture 13, was carried out according to the procedure
described in Chart 2. In particular, 2-amino-6-methylbenzoic
acid 9 was converted into the corresponding acid chloride by
reaction with SOCl2 and then reacted with the suitable anilines
to afford amides 10a,b. Quinazolinone scaffolds 12a,b were
then built up by further reaction of 10a,b with a 2-chloroacetyl
chloride derivatives 11a,b in refluxing acetic acid. Chlorides
12a,b were finally reacted with adenine in the presence of
K2CO3 to afford 4 and racemic 13. Single diasteroisomers 5−8
were isolated from 13 by preparative chiral HPLC using a
Chiralpak IA column (see Experimental Section for details).
Spectroscopic Characterization of Compound 1. First
of all, we investigated whether a commercial batch of
compound 1 consists of an achiral compound with free
rotation of the tolyl fragment, an optically pure atropisomer, or
a mixture of both atropisomers. For this purpose, compound 1
purchased from Symansis was fully characterized in CDCl3 by
means of NMR experiments. The 1D NMR spectrum
confirmed the proposed structure for 1 (Supporting
Information, Figure S1). As 1H NMR cannot detect the
existence of mirror-image enantiomers, the chiral reagent (1R)-
1-(anthracen-9-yl)-2,2,2-trifluoroethanol (R-Pirkle) was used to
generate diastereoisomeric complexes with compound 1 in
solution. This approach allowed identification enhanced shift
effects resulting from the formation of hydrogen bonds
between the ligand and the chiral reagent. Shift effects were
detected using a 1:10 molar ratio (Figure S2). In particular,
splitting of signals for the methylene group connecting the two
moieties and for the two methyl groups on the quinazolinone
nucleus and on the phenyl ring were clearly observed (Figure
2), consistent with the presence of two atropodiastereoisomers
in solution, indicating that the commercial batch of compound
1 is indeed a mixture of the aS (2) and aR (3) atropisomers.
Separation and Characterization of Atropisomers 2
and 3. The existence of atropisomer chirality for 1 was also
confirmed by HPLC on a chiral support, which shows the
existence of two distinct peaks (A and B). The first eluting
atropisomer [peak A, (−) compound] possesses the aS chiral
center (i.e., compound 2, Chart 1), while the second eluting
atropisomer [peak B, (+) compound 3, Chart 1] had an aR
chiral center, as deduced by vibrational circular dichroism
(VCD) analysis in combination with density functional theory
(DFT) calculations19 (see Experimental Section for details).
Having isolated and identified the structure of atropisomers 2
and 3, we next evaluated their thermal interconversion in
dimethyl sulfoxide (DMSO) at two different temperatures and
two concentrations by chiral HPLC/MS. At −20 °C, no
racemization was observed for either compound 2 or
compound 3 (data not shown). At 37 °C (Figure 3), both
Figure 3. Time course of reciprocal conversion of 2 (circles) to 3
(triangles) in dimethyl sulfoxide (DMSO), starting from concen-
trations of 1 mM (hollow) or 10 mM (filled). Percentages of original
atropisomer found in DMSO solution are reported as a function of
time.
compound 2 and compound 3 underwent a slow racemization
process, with nearly a 5% of conversion after 9 weeks of
incubation in solution, irrespective of the starting concentration
of the sample.
Altogether, these findings confirm that compound 2 and
compound 3 can be isolated as optically pure compounds and
that their stability is sufficient for further investigations on
single atropisomers.
Calculation of Atropisomer Interconversion Barrier.
The finding that compounds 2 and 3 showed a slow but
observable racemization in solution of dimethyl sulfoxide
(DMSO) prompted us to estimate the rate of atropisomer
conversion by means of computational methods. To this end,
we calculated the torsional energy barrier around the bond
connecting the quinazolin-4-one moiety with the ortho-tolyl
fragment in gas phase by means of quantum mechanics (QM)
and in explicit solvent using a hybrid quantum mechanics/
molecular mechanics (QM/MM) approach.20 The use of a
QM-based potential is necessary to obtain information about
the energy of the transition state (TS) separating the two stable
atropisomers. As classical force fields are not expected to give
robust energetic estimates for TS configurations unless they are
specifically parametrized for this purpose,21 DFT-based
approaches22 were used to generate TS structures and
D
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
subsequently to calculate the barrier for atropisomer con-
version.
applying an umbrella-sampling approach.26 Also in this case, the
conformational sampling around the bond connecting the
quinazolin-4-one moiety with the ortho-tolyl one was
performed from 90° to −90°. The shape of the resulting free
energy curve was similar to that calculated in gas phase for
model compound (Figure 5), whereas the free-energy barrier
for the conversion of 2 in 3 was 24.0 kcal/mol.
At first, the relationship between the potential energy and the
rotation around the ϕ dihedral angle in 2,5-dimethyl-3-(o-
tolyl)quinazolin-4(3H)-one, a simplified model of compound 2
(Figure 4), was assessed with the use of B3LYP hybrid
functional23 and the 6-31G* basis set.
As SCC-DFTB is a semiempirical method known to
underestimate energy barriers,22 the free energies of the
approximate transition state and of the reactant were corrected
at B3LYP/6-31G* level (see Experimental Section for details).
The corrected B3LYP/6-31G*//SCC-DFTB/AMBER free-
energy barrier was as high as 28.5 kcal/mol, very similar to
that found for the model compound in gas phase (28.0 kcal/
mol) and for similar quinazolinone inhibitors of PI3Kδ recently
reported in the literature.27 Thus, the rate of atropisomer
conversion seemed to primarily depend on the steric repulsion
between the quinazolin-4-one oxygen and the ortho-methyl
group of the tolyl fragment, whereas solvation and entropic
effects played a negligible role. The calculated barrier of 28.5
kcal/mol corresponds to a half-relaxation time for racemization
of nearly 90 days at 37 °C (deduced from the Eyring
equation),28 suggesting that compounds 2 and 3 can be isolated
and tested in vitro without racemization, in agreement with
results obtained from stability studies. According to the
classification system of atropisomers,8 compound 2 belongs
to class 2 as it experiences delayed axial interconversion, with a
half-relaxation time greater than hours.
Conformational Analyses of Compound 1 Derivatives
by Well-Tempered Metadynamics. We continued our
computational investigation on the conformational preferences
displayed by compound 2 by characterizing its conformational
FES in explicit solvent with the use of a well-tempered
metadynamics (wt-META-D) approach29 and OPLS2005 force
field.30 To this end, the two dihedral angles χ1 and χ2,
describing the possible orientations of the adenine ring with
respect to the quinazolin-4-one nucleus (Figure 4), were
employed as collective variables during a 50 ns long wt-META-
D simulation (see Experimental Section for details). Different
from molecular dynamics (MD) requiring long simulation time
to evaluate conformational changes, META-D has emerged as a
fast, robust, and efficient tool to evaluate free energy changes of
solvated systems.31 This approach is an artificial MD simulation
in which the evolution of the system is biased by a history-
dependent potential constructed as a sum of Gaussian functions
centered along a set of collective variables (CVs), which are
assumed to provide a description of the desired conformational
transformation.32 The added Gaussians allow the system to
escape from local minima33 and are fundamental to reconstruct
the conformational free energy surface (FES).34
Figure 4. Chemical formulas of the compounds employed in the
calculation of the rotational barrier around the ϕ dihedral angle
(depicted in red) by QM or QM/MM methods or for exploration of
the free-energy surface around χ1 and χ2 dihedral angles (depicted in
blue) by wt-metadynamics.
In the first torsional scan, rotation of the ϕ dihedral angle
was performed from 90° to −90°, with 0° corresponding to the
eclipsed conformation having the methyl group in contact with
the carbonyl oxygen. The ϕ angle was set to different values,
and a constrained geometry optimization was performed at
B3LYP/6-31G* level. The potential energy values thus
obtained are represented in Figure 5. On the basis of the line
connecting these values, the potential energy of atropisomer
conversion amounted to 28.0 kcal/mol.
Figure 5. Potential energy profile (blue line) in gas phase along the
dihedral angle ϕ for the model compound at B3LYP/6-31G* level and
free-energy profile (red line) in explicit water along the same dihedral
for compound 2 at SCC-DFTB/AMBER level of theory.
The resulting two-dimensional FES (Figure 6) shows the
presence of six distinct basins (A−F) corresponding to specific
conformations of compound 2. All these minima are featured
by a comparable free energy and therefore similarly populated
in solution. Calculations also showed that the conformation
assumed by compound 2 in the complex with PI3Kδ enzyme
(χ1 = 97° and χ2 = −30°, as obtained from the X-ray structure
depicted in Figure 1) corresponds to a well-defined free-energy
minimum (basin E). This finding suggests the accommodation
of 2 within PI3Kδ is driven by conformational selection; i.e.,
compound 2 is not forced into a high-energy state by the ATP
binding site of the enzyme.
The rotation of the same dihedral angle from −90° to 180°
(i.e., with the two methyl groups approaching each other) gave
higher energy values due to unfavorable steric interactions of
the methyl group of the tolyl substituent with the 2-methyl
group of the quinazolin-4-one scaffold (data not shown). This
is consistent with the results of recently reported calculations10
and points out that the passage of the ortho-methyl over the
carbonyl oxygen is more likely to occur during atropisomer
interconversion. Subsequently, we calculated the free energy of
atropisomer conversion for compound 2 in explicit water, using
a hybrid QM/MM potential (SCC-DFTB/AMBER)24,25 and
E
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 6. Two-dimensional plot of the conformational free energy surface of 2 (left panel) and 5−8 (right panels) in water constructed in the χ1 and
χ2 space from 50 ns long wt-META-D using OPLS-2005 force field. The contours are drawn at 1 kcal/mol. The triangle indicates χ1 and χ2 values
observed for compound 2 in the X-ray complex with PI3Kδ enzyme.
Subsequently, we investigated the conformational prefer-
ences of the diasteromeric couple formed by compounds 5 and
6 that, in addition to the aS chiral axis, also possess an S or an R
chiral center, respectively. Both compounds were submitted to
wt-META-D simulations in explicit water employing the same
computational protocol used for 2. The resulting FESs are also
reported in Figure 6. In both cases, the introduction of a methyl
group on the hinge methylene connecting the adenine ring to
the quinazolin-4-one nucleus drastically affects the energetics of
the conformational space along χ1 and χ2 compared to the
parent compound 2. In the case of the aS, S isomer (compound
5), A, B, and D free-energy areas become low-populated states
being separated by the global minimum C of nearly 4 kcal/mol
in case of D and more than 7 kcal/mol in the cases of A and B.
Conversely, basin E (corresponding to the bioactive con-
formation of 2) remained an accessible conformational state
being higher than C by less than 1 kcal/mol. In the case of the
aS, R isomer (compound 6), we obtained a specular FES, i.e.,
with basin D as a global minimum and C, E, and F regions as
high free-energy areas, separated from the former minimum by
approximately 3, 8, and 6 kcal/mol.
evaluation of conformational energies, the geometries of OPLS-
2005 minima identified with the META-D protocol were
optimized at DFT level using the M06-2X density functional.36
Geometry optimization (energy minimization with 6-31G**
basis set) were performed in implicit water, and the resulting
structures were utilized for single point energy calculations with
a larger basis set, cc-PVTZ(-f), using the same density
functional and solvation model. Considering that compounds
5 and 6 are enantiomers of compounds 8 and 7, respectively,
we limited these calculations to compounds 2, 5, and 6. M06-
2X/cc-PVTZ(-f) energies calculated for conformers A−F of
compounds 2, 5, and 6 are reported in Table S1 of the
Supporting Information. For compound 2, the energy differ-
ence between the minimum-energy conformation (conformer
A of the FES) and the PI3Kδ-binding minimum is 2.4 kcal/mol,
suggesting that its accommodation into the PI3Kδ binding site
involves significant strain (ΔEstrain). As for compound 5, the
ΔEstrain value is slightly smaller (1.7 kcal/mol), whereas for
compound 6 it increases up to 3.7 kcal/mol. The calculated
energy penalties for reaching the binding conformation mirror
the difference in inhibitory potency at PI3Kδ enzyme.
Similar FESs were obtained when the wt-META-D
simulations were performed for the aR, S (compound 7) and
aR, R (compound 8) diastereomeric pair (Figure 6), confirming
that the conformational space of these derivatives is driven
primarily by the spatial orientation of the newly introduced
methyl group, whereas the chiral axis lying on the ϕ dihedral
has a limited effect on the conformational free energy of the
identified basins. It is important to point out that the
conformational free-energy differences estimated for com-
pounds 2, 5−8 by wt-META-D refer to simulations that
reached convergence (see Experimental Section for details and
Figures S3−S7). wt-META-D simulations were replicated for
compounds 2, 5, and 6 applying different starting velocities,
similar to those reported in ref 35. No significant changes were
observed in the computed FESs calculated in three replicas
(Figure S8).
While the FESs point out that the lower inhibitor potency of
compound 6 with respect to its diastereoisomer 5 can be
attributed to an energy strain, resulting from lower abundance
of the active conformation, the difference of the free energy
calculated by metadynamics (approximately 7 kcal/mol)
appears to be overestimated. This may be due to the low
accuracy of the force-field functions when dealing with
nonminimum energy geometries. To get a more accurate
Biological Evaluation. Following these extensive analytical
and computational analyses, compounds 2 and 3 were next
tested for their ability to inhibit recombinant PI3Kδ enzyme
and to block the release of the inflammatory mediator
interleukin-6 (IL-6) from human peripheral blood mononuclear
cells (PBMCs).37 As shown in Table 1, compounds 2 and 3
have comparable potency in both these assays, indicating that
the orientation of the ortho-tolyl fragment is not relevant for
binding to the enzyme. This is in line with the visual inspection
of the crystal structure of the PI3Kδ−2 complex showing that
compound 3 could similarly fit the ATP-binding site of the
enzyme. This conclusion was further confirmed by compound
4, which lacks atropisomerism and exhibited a biological profile
similar to those of compounds 2 and 3.
On the contrary, the spatial orientation of the methyl group
introduced on the hinge connecting adenine and quinazolin-4-
one moieties considerably affects the inhibitory potency of the
synthesized atropo diastereomers 5−8. The aS, S isomer 5 was
considerably more potent than the aS, R isomer 6 which
registered a 30-fold drop in the inhibitory activity on PI3Kδ
enzyme. Similar findings were obtained comparing the activity
of the aR, S isomer (7) with the aR, R one (8), with the former
being nearly 10-fold more potent than the latter.
F
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Potency Values on Recombinant Human PI3Kδ
(Expressed as pKi Values) and on Interleukin-6 (IL-6)
Release from Human Blood Mononuclear Cells (Expressed
as pIC50 Values) for Compounds 2−8
CONCLUSIONS
■
Atropisomerism is the drug chirality due to slowly interconvert-
ing compound conformations as a result of either steric or
electronic hindrance allowing their isolation. Compounds
exhibiting this chirality can have different efficacy, selectivity,
and pharmacokinetic profile similar to enantiomers. Even
though recent literature has contributed to increase the
awareness of their potential issues, they are frequently
overlooked in medicinal chemistry projects. Besides, the lack
of clear guidance from the regulatory agencies further
complicates this scenario.
Here we have demonstrated with the use of NMR chiral
reagents that compound 1, a standard PI3Kδ inhibitor
extensively used to carry out in vitro and in vivo experiments
for this enzyme, is a mixture of atropisomers. Besides, we have
shown by using HPLC that at 37 °C they undergo a very slow
racemization process, consistent with their calculated free
energy conversion of 28.5 kcal/mol in explicit water so that
they can be separated and characterized. The QM/MM
calculations have shown that the rate of atropisomer conversion
mainly depends on the steric repulsion between the quinazolin-
4-one oxygen and the ortho-methyl group, whereas solvation
and entropic effects seem to be negligible.
The introduction of a stereogenic center on top of the axial
chirality changes the biological profiles of the diastereisomers
obtained. In particular, the S stereogenic center increases both
the enzymatic and the cell-based potency inhibition with
respect to the R one. Besides, the combination of the S
stereogenic center with the aS chirality provides the most
potent compound. This is also the only derivative out of the
four diastereoisomers showing that the bioactive conformation
observed in the crystal structure is energetically accessible.
a
b
c
compd
chirality
PI3Kδ, pKi
IL-6 release, pIC50
2
3
4
5
6
7
8
aS
7.9 0.03
7.7 0.1
7.6 0.1
8.5 0.1
6.9 0.04
8.8 0.1
7.9 0.03
7.6 0.1
7.3 0.1
7.3 0.4
8.2 0.5
aR
aS, S
aS, R
aR, S
aR, R
d
NT
8.9 0.1
d
NT
a
Chirality was determined with vibrational circular dichroism (VCD;
b
see Experimental Section for details). pKi values were obtained with
Cheng and Prusoff equation from experimental IC50 values and are the
c
mean
SD of four determinations performed in duplicate. pIC50
values are the mean
performed in quadruplicate. Not tested.
SD of two independent determinations
d
Visual inspection of the PI3Kδ−2 complex (Figure 1)
suggests that the introduction of a small group on the
methylene hinge of 2 should not affect the inhibitory potency
of the class, as both the pro-S and pro-R hydrogens of this
methylene point toward an empty region of the ATP binding
site of PI3Kδ. On the basis of the simulations reported above,
we propose that compound 5 inhibits PI3Kδ more efficiently
than 6 because it can easily adopt the conformation assumed by
the reference inhibitor 2 in the crystal structure with PI3Kδ
(Figure 7a), while compound 6 can assume the bioactive
EXPERIMENTAL SECTION
■
Chemical Synthesis. General Methods. All reactions were carried
out under argon atmosphere. Solvents were purchased from
commercial sources and used without further drying: ammonium
formate for HPLC, ≥99.0% (Fluka); acetonitrile LC−MS Ultra
CHROMASOLV tested for UHPLC−MS (Fluka); methanol LC−MS
Ultra CHROMASOLV tested for UHPLC−MS (Fluka).
Compound 1 (2-((6-amino-9H-purin-9-yl)methyl)-5-methyl-3-(o-
tolyl)quinazolin-4(3H)-one) was purchased from Symansis (www.
symansis.com). Chiral reagent (1R)-1-(anthracen-9-yl)-2,2,2-trifluor-
oethanol (R-Pirkle) was purchased from Sigma. Solvents DMSO-d6
and CDCl3 were 99.96% pure and were purchased from Merck. Silica
gel chromatography purifications were performed in flash conditions
using either glass columns packed with 230−400 mesh silica gel
(Merck) or prepacked silica gel cartridges (Biotage). Analytical thin
layer chromatography (TLC) was carried out on Merck silica gel plates
(silica gel 60 F254). 1H NMR and 13C NMR spectra were recorded on
a Bruker Avance 200 spectrometer, using CDCl3 or DMSO-d6 as
solvent. Chemical shifts (δ scale) are reported in parts per million
(ppm) relative to the central peak of the solvent. Coupling constants
(J values) are given in hertz (Hz). EI-MS spectra (70 eV) were taken
on a Fison Trio 1000; molecular ions (M+ or M−) are given. ESI-MS
spectra were taken on a Waters Micromass ZQ instrument; only
molecular ions (M + 1 or M − 1) are given. Melting points were
determined on a Buchi SMP-510 capillary melting point apparatus and
are uncorrected. Microwave reactions were carried out using Biotage
Initiator microwave synthesizer. Specific rotations were measured on a
Bellingham Stanley ADP440+ polarimeter. The purity of the
compounds reported in the manuscript was established through
HPLC methodology, with both UV and MS detection. All the tested
compounds reported in the manuscript have a chemical purity greater
than 95%.
Figure 7. (a) Superposition of a minimum free-energy conformation
of compound 5 (yellow carbon atoms, taken from energy basin E of
Figure 6) on the bioactive one assumed by compound 2 (green carbon
atoms) in its crystallographic complex with PI3Kδ enzyme. (b−d)
Superposition of highly populated free-energy conformations of
compound 6 (orange carbon atoms, taken from basins A, B, and D
of Figure 6) on the bioactive conformation of compound 2 (green
carbon atoms).
conformation only paying a free-energy penalty. Moreover, not
one of the other accessible conformations identified for 6 (i.e.,
lying on basins A, B, and D of Figure 6) is able to fit the
pharmacophore elements possessed by compound 2 in its
complex with PI3Kδ as indicated by the superpositions
reported in Figure 7b−d.
G
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(aS)-2-[(6-Amino-9H-purin-9-yl)methyl]-5-methyl-3-(2-
methylphenyl)quinazolin-4(3H)-one (2). IC78114 was separated by
preparative chiral HPLC on a Chiralpak IA under isocratic elution at
60 mL/min eluting with DCM + 0.1% diethylamine to give as first
7.04 (m, 1 H), 5.19−5.47 (m, 1 H), 2.73 (s, 3 H), 2.13 (s, 3 H), 1.59−
1.74 (m, 3 H).
2-Amino-6-methyl-N-phenylbenzamide (10a). Thionyl chloride
(0.290 mL, 3.97 mmol) was added to a stirred solution of 2-amino-6-
methylbenzoic acid (0.2 g, 1.32 mmol) in toluene (20 mL), and the
mixture was refluxed for 18 h. The reaction was then cooled to room
temperature, and the solvent was removed under reduced pressure and
stripped down twice with toluene (25 mL). The residue was dissolved
in DCM (20 mL) and reacted with TEA (0.277 mL, 1.98 mmol) and
aniline (0.181 mL, 1.98 mmol). The slurry was then refluxed for 2 h.
The reaction mixture was purified with a silica gel cartridge eluting
with DCM to give the title compound 10a (0.120 g, 0.53 mmol, 40%
20
eluting isomer 2 (tR = 10.9 min). MS (ESI): 398.1 [M + H]+; [α]D
+40.6 (c 2.4, CHCl3). 1H NMR (400 MHz, DMSO-d6): δ ppm 2.18 (s,
3 H), 2.73 (s, 3 H), 4.56−5.23 (m, 2 H), 7.07−7.25 (m, 3 H), 7.30 (d,
1 H, J = 7.1 Hz), 7.39−7.54 (m, 4 H), 7.61 (t, 1 H, J = 7.9 Hz), 8.05
(d, 2 H, J = 4.4 Hz).
(aR)-2-[(6-Amino-9H-purin-9-yl)methyl]-5-methyl-3-(2-
methylphenyl)quinazolin-4(3H)-one (3). Staring from the commer-
cial bath of compound 1, atropisomer 3 was recovered as second
eluting compound (tR = 12.8 min) in the same chromatographic run
employed to isolate 2. MS (ESI): 398.1 [M + H]+; [α]D20 −39.7 (c 2.6,
1
yield) as colorless solid. MS (ESI): 330−332 [M + H]+. H NMR
(CDCl3): δ 3.42 (s, 3 H), 3.49 (s, 3 H), 4.11 (s, 3 H), 6.19 (s, 1 H),
6.90 (d, 1 H, J = 16.2 Hz), 7.08 (d, 1 H, J = 16.2 Hz), 7.29−7.37 (m, 3
H), 7.50 (s, 1 H). 13C NMR (CDCl3): δ 27.9, 31.7, 31.8, 90.9, 111.4,
116.1, 124.9, 126.3, 128.5, 130.1, 131.4, 134.9, 135.6, 138.1, 139.5,
151.5, 155.8. FTIR (Nujol, cm−1): 1647, 1685, 2854, 2924. Mp: 200
°C.
1
CHCl3). H NMR (400 MHz, DMSO-d6): δ 2.18 (s, 3 H), 2.73 (s, 3
H), 4.56−5.23 (m, 2 H), 7.07−7.25 (m, 3 H), 7.30 (d, 1 H, J = 7.1
Hz), 7.39−7.54 (m, 4 H), 7.61 (t, 1 H, J = 7.9 Hz), 8.05 (d, 2 H, J =
4.4 Hz).
2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylquinazo-
lin-4(3H)-one (4). 2-(Chloromethyl)-5-methyl-3-phenylquinazolin-
4(3H)-one 12b (0.052 g, 0.183 mmol) was dissolved in acetonitrile
(10 mL) and few drops of DMF. K2CO3 (0.038 g, 0.274 mmol) and
adenine (0.037 g, 0.274 mmol) were added, and the mixture was
reacted at 65 °C for 6 h. Acetonitrile was removed under reduced
pressure, and the suspension was poured into water (30 mL) and
stirred at room temperature for 30 min and then cooled to 5 °C. The
solid was filtered and dried and the resulting material was purified by
silica gel flash chromatography with a gradient of DCM and MeOH to
give the title compound 4 (0.037 g, 0.097 mmol, 53% yield). MS
(ESI): [M + H]+ 313.2. 1H NMR (300 MHz, DMSO-d6) δ 7.98−8.11
(m, 2 H), 7.41−7.69 (m, 3 H), 7.19−7.33 (m, 2 H), 7.10−7.18 (m, 1
H), 5.00 (s, 2 H), 2.73 (s, 3 H).
(aS,S)-2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylqui-
nazolin-4(3H)-one (5). Mixture of diastereoisomers 13 was separated
by chiral HPLC on a Chiralpak IC under isocratic elution at 0.8 mL/
min (phase A/B = 65/35% v/v; phase A, hexane; phase B, EtOH +
0.1% isopropylamine) to give 5 as the first eluting isomer (tR = 11.2
min). 1H NMR (400 MHz, DMSO-d6) δ 8.26 (s, 1 H), 7.96 (s, 1 H),
7.65−7.76 (m, 1 H), 7.55−7.63 (m, 1 H), 7.41−7.52 (m, 3 H), 7.29−
7.38 (m, 2 H), 7.19 (bs, 2 H), 5.25−5.43 (m, 1 H), 2.72 (s, 3 H), 1.89
(s, 3 H), 1.78 (d, J = 7.06 Hz, 3 H).
(aS,R)-2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylqui-
nazolin-4(3H)-one (6),. Mixture of diastereoisomers 13 was separated
by chiral HPLC on a Chiralpak IC under isocratic elution at 0.8 mL/
min (phase A/B = 65/35% v/v; phase A, hexane; phase B, EtOH +
0.1% isopropylamine), giving 6 as the second eluting isomer (tR = 12.6
min). 1H NMR (400 MHz, DMSO-d6) δ 8.16 (s, 1 H), 7.94 (s, 1 H),
7.60−7.77 (m, 1 H), 7.42−7.54 (m, 2 H), 7.26−7.38 (m, 2 H), 7.16
(bs, 2 H), 7.03−7.09 (m, 1 H), 6.73−7.01 (m, 1 H), 5.15−5.39 (m, 1
H), 2.69 (s, 3 H), 2.10 (s, 3 H), 1.64 (d, J = 6.62 Hz, 3 H).
(aR,S)-2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylqui-
nazolin-4(3H)-one (7). Isomer 7 coeluted with 8 with the first chiral
cromatography employed to isolate 5 and 6. The title compound was
thus isolated with another chiral HPLC on Chiralpak AD-H under
isocratic elution at 0.8 mL/min (phase A/B = 60/40% v/v; phase A,
hexane; phase B, EtOH + 0.1% isopropylamine) which gave
compound 7 as first eluting isomer (rt 9.4 min). 1H NMR (400
MHz, DMSO-d6) δ 8.26 (s, 1 H), 7.96 (s, 1 H), 7.65−7.75 (m, 1 H),
7.53−7.63 (m, 1 H), 7.39−7.51 (m, 3 H), 7.28−7.38 (m, 2 H), 7.19
(bs, 2 H), 5.21−5.45 (m, 1 H), 2.72 (s, 3 H), 1.89 (s, 3 H), 1.78 (d, J
= 7.06 Hz, 3 H).
2-Amino-6-methyl-N-o-tolylbenzamide (10b). Thionyl chloride
(0.579 mL, 7.94 mmol) was added to a stirred solution of 2-amino-6-
methylbenzoic acid (0.4 g, 2.65 mmol) in toluene (30 mL), and the
mixture was refluxed for 18 h. The reaction was then cooled to room
temperature, and the solvent was removed under reduced pressure and
stripped down twice with toluene. The residue was dissolved in DCM
(30 mL) and reacted with TEA (0.553 mL, 3.97 mmol) and o-
toluidine (0.424 mL, 3.97 mmol). The slurry was then refluxed for 2 h.
The reaction mixture was purified by silica gel flash cromatography,
eluting with a gradient of petroleum ether and EtOAc to give the title
compound 10b (0.278 g, 1.16 mmol, 44% yield) as pale yellow gum.
MS (ESI): [M + H]+ 241.4. 1H NMR (300 MHz, CDCl3): δ 2.29 (s, 3
H); 2.44 (s, 3 H); 4.16 (bs, 2 H); 6.53−6.56 (d, 1 H); 6.62−6.64 (s, 1
H); 7.06−7.11 (t, 1 H); 7.13−7.18 (t, 1 H); 7.23−7.26 (m, 2 H); 7.65
(bs, 1 H); 7.84−7.87 (s, 1 H). 13C NMR (75.5 MHz, CDCl3): δ 18.16;
20.34; 114.08; 120.37; 122.98; 123.71; 125.75; 126.76; 130.15; 130.27;
130.72; 135.41; 135.47; 144.97; 167.92.
2-(Chloromethyl)-5-methyl-3-phenylquinazolin-4(3H)-one (12a).
2-Amino-6-methyl-N-phenylbenzamide 10a (0.120 g, 0.53 mmol) was
dissolved in AcOH (5 mL) and reacted with 2-chloroacetyl chloride
11a (0.084 mL, 1.06 mmol), and the mixture was refluxed for 18 h.
The reaction was then cooled and concentrated under reduced
pressure, and the residue was straightforward purified by silica gel flash
cromatography with a gradient of petroleum ether and AcOEt to give
the title compound 12a (0.052 g, 0.183 mmol, 34% yield). MS (ESI):
1
[M + H]+ 285.2. H NMR (300 MHz, CDCl3): δ 2.86 (s, 3 H); 4.26
(s, 2 H); 7.31−7.33 (dd, 1 H); 7.33−7.40 (m, 2 H); 7.56−7.70 (m, 5
H). 13C NMR (75.5 MHz, CDCl3): δ 23.01; 43.56; 120.11; 125.97;
128.85 (2 C); 129.69; 129.87 (2 C); 130.64; 133.95; 136.21; 141.76;
148.59; 151.18; 162.66.
2-(1-Chloroethyl)-5-methyl-3-o-tolylquinazolin-4(3H)-one (12b).
2-Amino-6-methyl-N-o-tolylbenzamide 10b (0.278 g, 1.16 mmol) was
dissolved in AcOH (10 mL) and reacted with 2-chloropropanoyl
chloride 11b (0.225 mL, 2.31 mmol) and the mixture refluxed for 18
h. The reaction was cooled and concentrated under reduced pressure
and the residue was straightforward purified by silica gel flash
chromatography with a gradient of petroleum ether and AcOEt to give
the title compound 12b (0.082 g, 0.262 mmol, 23% yield) as yellow
1
gum. MS (ESI): [M + H]+ 313.2. H NMR (400 MHz, CDCl3): δ
1.87−1.88 (d, 3 H); 1.90−1.91 (d, 3 H); 2.09 (s, 3 H); 2.29 (s, 3 H);
2.87 (s, 6 H); 4.34−4.39 (q, 1 H); 4.65−4.70 (q, 1 H); 7.09−7.11 (d,
1 H); 7.31−7.32 (d, 2 H); 7.36−7.49 (m, 7 H); 7.65−7.70 (m, 4 H).
13C NMR (100.6 MHz, CDCl3): δ 17.65; 18.10; 22.28; 22.53; 23.03;
(aR,R)-2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylqui-
nazolin-4(3H)-one (8). Compound 8 was separated from 7 under
chiral HPLC on Chiralpak AD-H under isocratic elution at 0.8 mL/
min (phase A/B = 60/40% v/v; phase A, hexane; phase B, EtOH +
0.1% isopropylamine). Compound 8 was indentified as second eluting
isomer compound 8 (tR = 16.4 min). 1H NMR (400 MHz, DMSO-d6)
δ 8.19 (s, 1 H), 7.98 (s, 1 H), 7.62−7.75 (m, 1 H), 7.42−7.56 (m, 2
H), 7.31−7.40 (m, 2 H), 7.19 (bs, 2 H), 7.05−7.13 (m, 1 H), 6.93−
23.06; 52.45; 52.57; 119.64; 119.91; 126.16; 126.21; 127.33; 127.59;
127.92; 129.67; 129.79; 129.86; 130.34; 131.17; 131.89; 133.76;
135.20; 135.27; 135.54; 137.49; 141.73; 148.78; 148.81; 154.06;
154.23; 161.92; 162.19.
2-((6-Amino-9H-purin-9-yl)methyl)-5-methyl-3-phenylquinazo-
lin-4(3H)-one (13). 2-(1-Chloroethyl)-5-methyl-3-o-tolylquinazolin-
4(3H)-one 12b (0.057 g, 0.182 mmol) was dissolved in acetonitrile
(15 mL) with few drops of DMF. K2CO3 (0.038 g, 0.273 mmol) and
H
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
adenine (0.037 g, 0.273 mmol) were added, and the mixture was
reacted at 65 °C for 24 h. Solvent was then removed, and the reaction
was taken up with DMF (10 mL), and then the mixture was heated
under microwave irradiation at 130 °C for 4 h. The crude was directly
purified by preparative HPLC (Shimadzu, neutral phase) to give the
title compound 13 (0.020 g, 0.049 mmol, 27% yield).
NMR Characterization of Compound 1 and 1:R-Pirkle
Complex. Compound 1. The title compound was dissolved in 0.75
mL of CDCl3 99.96% at room temperature. The obtained solution was
transferred into an NMR tube and taken to the NMR analysis.
Complete characterization was run in CDCl3 99.96%. Compound 1
(CDCl3): 1H NMR (400 MHz, chloroform-d) δ 8.26 (s, 1H), 7.89 (s,
1H), 7.54−7.63 (m, 1H), 7.45 (d, J = 4.41 Hz, 2H), 7.33−7.42 (m,
2H), 7.26 (m, 1H), 7.21 (d, J = 7.50 Hz, 1H), 5.66 (br s, 2H), 4.75−
5.18 (m, 2H), 2.83 (s, 3H), 2.22 (s, 3H).
imposing constraining potentials along the ϕ dihedral angle of 2
(Figure 4), with a force constant of 500 kcal mol−1 Å−2 from the value
of 90° up to −90°. In this way, the ortho methyl group of 2 was
gradually moved from one side of the plane identified by the
quinazolin-4-one nucleus to the other. The interval between each
harmonic potential was 10° for a total of 19 simulation windows. The
sampling for each harmonic potential consisted of 200 ps, for a total of
3.8 ns of QM/MM simulation. We discarded the statistics from the
first 100 ps to allow a suitable relaxation from the imposed bias. The
final free energy profile was obtained by using the weighted histogram
analysis method (WHAM)42 on the collected statistics.
High-Level Correction. The geometries of the starting structure and
of the approximate TS from SCC-DFTB/AMBER simulations were
isolated, and gas-phase energy calculations were performed at the
SCC-DFTB and B3LYP/6-31G* levels. The corrected energies were
calculated applying the computational protocol described in ref 43. In
brief the revised energies were obtained by subtracting from the total
QM/MM energy the SCC-DFTB energy of the isolated QM region
and adding the B3LYP energy. The B3LYP corrected free energies
therefore consist of the B3LYP vacuum energy, of the AMBER MM
energy, and of the SCC-DFTB/AMBER99 QM/MM interaction
energy.
Conformational Analysis by wt-Metadynamics. wt-META-D
simulations were conducted with Desmond version 3.0.44 The dihedral
angles χ1 and χ2 reported in Figure 4 were used as collective variables
(CVs) to sample the conformational space of compounds 2 and 5−8.
Compounds 5−8 were built using the sketching tool of Maestro. The
geometry of the compounds was optimized using OPLS2005 force
field in combination with GB/SA water model. Then, the compounds
were placed into a cubic box of TIP3P water molecules,45 leaving a
minimum distance of 24 Å between any ligand atom and the edges of
the box. The solvated systems were first relaxed using two energy
minimization processes, aimed at relieving major steric clashes. While
in the first minimization stage all solute heavy atoms were constrained
with a force constant of 50 kcal mol−1 Å−2 and the convergence
threshold was set to 50 kcal mol−1, in the second minimization step
the threshold was decreased to 5 kcal mol−1 and all constraints were
removed. A plain MD simulation was then performed to relax each
system before the META-D run. The relaxation phase comprised (i)
100 ps in the NVT ensemble at 10 K with a constraint of 50 kcal/(mol·
Å2) on solute heavy atoms, (ii) 100 ps in the NPT ensemble at 10 K
with a constraint of 50 kcal/(mol·Å2) on solute heavy atoms, (iii) 100
ps in the NPT ensemble at 300 K with a constraint of 50 kcal/(mol·
Å2) on solute heavy atoms, (iv) 200 ps in the NPT ensemble at 300 K
without constraints. META-D simulations were conducted in the NPT
ensemble at 300 K and 1 atm for 50 ns, applying the Langevin
thermostat and barostat. The M-SHAKE46 was used to constrain all
bond lengths to hydrogen atoms. Long-range electrostatic interactions
were computed using the smooth particle mesh Ewald method,47
whereas short-range electrostatic interactions were cut off at 9 Å. A
RESPA integrator48 was used to compute the bonded, van der Waals,
and short-range electrostatic interactions every 2 fs, whereas long-
range electrostatics were computed every 6 fs. Gaussian potentials
were added every 0.09 ps with a height of 0.03 kcal/mol and a width of
5°. The OPLS2005 force field was used in all calculations. FESs were
then reconstructed from the Gaussian potentials using the Desmond
metadynamics analysis tool implemented in Maestro 9.2. The
convergence of wt-META-D simulations was assessed by evaluating
the time evolution of the relative free-energy levels of the key free-
energy minima. To this end, wt-META-D were run for 50 ns, while
FESs were reconstructed at every nanosecond of simulation. The
simulations were considered at convergence when the free-energy
difference among global and local minima became approximately
constant. For the compounds 2 and 5−8, simulations converged
approximately after 30 ns (see Supporting Information Figures S3−
S7). The FESs reported in Figure 6 are those reconstructed at the end
of the simulations.
Association Complex. Compound 1 (1 mg, 2.5 × 10−3 mmol) was
dissolved in 0.75 mL of CDCl3 99.96% at rt, and this solution was
added to chiral reagent (1R)-1-(anthracen-9-yl)-2,2,2-trifluoroethanol
(R-Pirkle, 6.9 mg, 25 × 10−3 mmol, molar ratio 1:10 with respect to
the standard). The mixture was transferred into an NMR tube and
taken to the NMR analysis. Compound 1:R-Pirkle (CDCl3): 1H NMR
(400 MHz, chloroform-d) δ 8.97 (br s, 9H), 8.53 (s, 11H), 7.89−8.27
(m, 31H), 7.44−7.69 (m, 44H), 7.29−7.40 (m, 3H), 7.24−7.26 (m,
2H), 6.89−7.12 (m, 1H), 6.65 (q, J = 7.83 Hz, 10H), 5.49 (br s, 2H),
4.42−4.99 (m, 2H), 3.36 (br s, 10H), 2.81 (d, J = 7.83 Hz, 3H), 2.11
(d, J = 14.09 Hz, 3H).
Stability Study. Compound 2 and compound 3 were dissolved in
dimethyl sulfoxide (DMSO) at 5 mg/mL (10 mM) and 0.5 mg/mL (1
mM) and stored at two different temperatures, i.e., 37 and −20 °C for
9 weeks. Samples were diluted to 1 μg/mL with methanol before
injection. The stability of both solutions was evaluated by HPLC/MS
using a Acquity UPLC system (Waters) coupled with a 5500-QTrap
MS instrument (AB Sciex) equipped with an ESI source. A column
Lux Cellulose-1, 5 μm, 4.6 mm × 250 mm, was employed at 30 °C.
Solutions used as eluent: HCOONH4 0.025 M, pH 3 (A); ACN/
MeOH 40/60 (B) (isocratic, 35% A, 65% B). The injection volume
was 5 μL, the flow was 0.6 mL/min, while the run time was 18 min. All
the MS data were acquired in positive-ion mode applying the multiple-
reaction-monitoring (MRM) settings.
Molecular Modeling. Potential Energy Barrier Estimation by
DFT. Quantum chemical calculations were performed using the Jaguar
7.838 within Maestro suite 9.2.39 In Jaguar, relaxed torsion scans of the
dihedral angle N−C1−C2−C3 of the model compound were defined
for a range of 180° at 15° intervals, starting from the value of 90°. In
the vicinity of the approximate transition state (at 0°) additional angle
values were collected (i.e., 10° and 5°). At each restrained torsion
angle, the geometry was optimized by using the hybrid functional
B3LYP in combination with B3LYP/6-31G* basis set. SCF and
convergence criteria were set at 5.0 × 10−5 hartrees or rms density
matrix change of 5.0 × 10−5 hartrees, whichever was satisfied first.
Model Building and Application of a QM/MM Potential.
Compound 2 was taken from the coordinate of 2X38.pdb and
imported into Maestro 9.2.34 Atom types and bond orders were
corrected, and hydrogen atoms were added. The resulting structure
was immersed in a box of TIP3P water molecules (1397 residues)
using the Xleap tool of AMBER 10.40 The total system size amounted
to 4240 atoms (box size of 37.1 Å × 41.0 Å × 39.6 Å). The system was
minimized using the AMBER99SB force field,41 in the SANDER
module of AMBER 10 and then used for QM/MM calculations. The
2−TIP3P system was thermalized (300 K) in the NVT ensemble for
100 ps using the QM/MM scheme implemented in the SANDER
module of the AMBER 10 suite. All the atoms of 2 (49 in total) were
treated with the self-consistent charge-density functional tight-binding
(SCC-DFTB) approach, as implemented in SANDER. The solvent
atoms were described with AMBER99SB force field. During QM/MM
MD simulations (either plain or biased), all the atoms of the system
(including hydrogens) were free to move, and a time step of 0.2 fs was
used to integrate the equation of motion.
Conformational Energies by DFT Calculations. OPLS-2005
minima identified on the FES of Figure 6 for compounds 2, 5, and
6 were used as input structures for DFT optimization in implicit
Umbrella Sampling Simulations. Starting from the equilibrated 2−
TIP3P system, an umbrella sampling calculation was performed by
I
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
solvent (PCM, water) at M06-2X/6-31G**/PCM = H2O level.
Solution energies were calculated using the Poisson−Boltzmann
equation solver implemented in Jaguar.49 The resulting geometries
were used as input structures for single point calculation at M06-2X/
cc-PVTZ(-f)/PCM = H2O level. Default convergence criteria for SCF
calculation and geometry optimization were applied.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.7b00247.
■
S
Vibrational Circular Dichroism (VCD) Studies. The absolute
configuration of each atropisomer of compound 1 (i.e., compounds 2
and 3) as well as of compounds 5−8 was determined recording VCD
spectra in tandem with calculations of the predicted spectra.50 In brief,
the experimental VCD spectra of the inhibitors were recorded on a
BioTools/Bomem Dual PEM chiral IR spectrometer (Biotools, Inc.,
Jupiter, FL) using a 100 μm liquid cell equipped with BaF2 windows.
Spectra were recorded using CDCl3 as solvent. For all samples, 40 000
scans were recorded at 4 cm−1 resolution and averaged, while the
spectrum obtained for the pure solvent was used for solvent
subtraction. Concentrations of 5 mg in 150 μL were used. A
systematic conformational analysis of 2−8 was carried out using the
MMFF94 molecular mechanics force51 field implemented in Sybyl-X
software.52 For each compound, potential energy minima obtained
were optimized using DFT at wB97XD/cc-pvtz level using a PCM
continuum model for solvent representation (chloroform). Boltzmann
weighted VCD spectra were then calculated using Gaussian 09
program53 at the same level of theory. Comparison of the DFT
calculated spectra to the experimental ones enables the absolute
configuration of 2, 3, and 5−8 to be reliably determined.54,55
Molecular formula strings and some data for compounds
2−8 (CSV)
1H NMR spectrum (CDCl3, 25 °C) of compound 1
alone or in complex with R-Pirkle; plots reporting time
evolution of the free-energy differences among key
minima identified on conformational FES reconstructed
by wt-META-D for compounds 2, 5−8; M06-2X/cc-
PVTZ(-f) energies for compounds 2, 5, and 6;
experimental details for the HPLC-based stability studies
performed on compounds 2 and 3 (PDF)
AUTHOR INFORMATION
Corresponding Authors
*M.M.: phone, +39 0521 905059; fax, + 39 0521 905006; e-
mail, marco.mor@unipr.it.
*A.M.C.: phone, +39 0521 279188; fax, + 39 0521 279880; e-
mail, am.capelli@chiesi.com.
■
Biological Test Methods. In Vitro Kinase Assay. Human
recombinant PI3Kδ was purchased from Millipore Ltd. (Billerica,
MA). Compounds were dissolved at 0.5 mM in DMSO and were
tested at different concentrations for their activity against PI3Kδ using
the ADP-Glo kinase assay (Promega, Madison, WI) according to the
manufacturer’s instructions. Briefly, the kinase reactions were
performed in 384-well white plates (Greiner Bio-One GmbH,
Frickenhausen). Each well was loaded with 0.1 μL of test compounds
and 2.5 μL of 2× reaction buffer (40 mM Tris, pH 7.5, 0.5 mM EGTA,
0.5 mM Na3VO4, 5 mM β-glycerophosphate, 0.1 mg/mL BSA, 1 mM
DTT), containing 50 μM PI and PS substrates (L-α-phosphatidylino-
sitol sodium salt and L-α-phosphatidyl-L-serine, Sigma-Aldrich, St.
Louis, MO) and the PI3K recombinant protein (PI3Kδ 1 ng/μL). The
reactions were started by adding 2.5 μL of 2× ATP solution to each
well (final concentration ATP 80 μM) and incubated for 60 min at
room temperature. Subsequently, each kinase reaction was incubated
for 40 min with 5 μL of ADP-Glo reagent, allowing depletion of
unconsumed ATP. Then, the kinase detection reagent (10 μL) was
added in each well to convert ADP to ATP and to allow the newly
synthesized ATP to be measured using a luciferase/luciferin reaction.
Following a 60 min incubation, the luminescence signal was measured
using a Wallac EnVision multilabel reader (PerkinElmer, Waltham,
MA). Curve fitting and pIC50 calculation were carried out using a four-
parameter logistic model in XLfit (IDBS, Guilford, U.K.) for Microsoft
Excel (Microsoft, Redmont, WA).
ORCID
Alessio Lodola: 0000-0002-8675-1002
Marco Mor: 0000-0003-0199-1849
Daniele Pala: 0000-0001-6160-6127
Silvia Rivara: 0000-0001-8058-4250
Present Address
∥D.P.: Chiesi Farmaceutici S.p.A., Largo Belloli, 11/A, Parma,
Italy.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS USED
■
PI3K, phosphoinositide 3-kinase; IL-6, interleukin-6; DFT,
density functional theory; FES, free-energy surface; wt-META-
D, well-tempered metadynamics; PBMC, peripheral blood
mononuclear cell; VCD, vibrational circular dichroism
Determination of PI3K Inhibitory Activity in the PBMCs Assay.
Human peripheral blood mononuclear cells (PBMCs) were purchased
from Lonza (Basel, CH), washed, and resuspended in RPMI 1640
medium (without phenol red) supplemented with 10% FBS, 2 mM
glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin
(Thermo Fisher Scientific, Waltham, MA). PBMCs were plated at a
density of 105 cells/well in 96-well plates coated with 6 μg/mL anti-
human CD3 antibody (Biolegend, San Diego, CA). Cells were treated
in RPMI 1640 medium (without phenol red) supplemented with 10%
FBS with different concentrations of PI3K inhibitors (10−12−10−5 M,
final DMSO concentration 0.2%), co-stimulated with 3 μg/mL anti-
human CD28 antibody (BD Biosciences, San Jose, CA), and incubated
for 72 h in an atmosphere of 95% air and 5% CO2 at 37 °C. Human
IL-6 was measured in the supernatants using paired antibody
quantitative ELISA kit (from Thermo Fisher Scientific, Waltham,
MA) according to the manufacturer’s instructions. pIC50 values were
determined from concentration−response curves by nonlinear
regression analysis using Graph Pad Prism, version 6 (GraphPad
Software, La Jolla, CA).
REFERENCES
■
(1) Oki, M. Recent Advances in Atropisomerism. In Topics in
Stereochemistry; Allinger, N. L., Eliel, N. L., Wilen, S. H., Eds.; John
Wiley & Sons: New York, 1983; Vol. 14, pp 1− 81, DOI: 10.1002/
9780470147238.ch1.
(2) Eichelbaum, M.; Gross, A. S. Stereochemical aspects of drug
action and disposition. Adv. Drug Res. 1996, 28, 1−6.
(3) Clayden, J.; Moran, W. J.; Edwards, P. J.; LaPlante, S. R. The
challenge of atropisomerism in drug discovery. Angew. Chem., Int. Ed.
2009, 48, 6398−6401.
(4) LaPlante, S. R.; Forgione, P.; Boucher, C.; Coulombe, R.; Gillard,
J.; Hucke, O.; Jakalian, A.; Joly, M.-A.; Kukolj, G.; Lemke, C.;
McCollum, R.; Titolo, S.; Beaulieu, P. L.; Stammers, T. Enantiomeric
atropisomers inhibit HCV polymerase and/or HIV matrix: character-
izing hindered bond rotations and target selectivity. J. Med. Chem.
2014, 57, 1944−1951.
(5) Webster, R.; Cole, S.; Gedge, J.; Roffey, S.; Walker, D.; Wild, W.
Pharmacokinetics and disposition of a novel NMDA glycine site
J
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
antagonist (UK-240,455) in rats, dogs and man. Xenobiotica 2003, 33,
541−560.
(c) Lodola, A.; De Vivo, M. The increasing role of QM/MM in
drug discovery. Adv. Protein Chem. Struct. Biol. 2012, 87, 337−362.
́
́
(21) Madarasz, A.; Berta, D.; Paton, R. S. Development of a true
(6) Elleraas, J.; Ewanicki, J.; Johnson, T. W.; Sach, N. W.; Collins, M.
R.; Richardson, P. F. Conformational studies and atropisomerism
kinetics of the ALK clinical candidate lorlatinib (PF-06463922) and
desmethyl congeners. Angew. Chem., Int. Ed. 2016, 55, 3590−3595.
(7) Glunz, W. P.; Mueller, L.; Cheney, L. D.; Ladziata, V.; Zou, Y.;
Wurtz, N. R.; Wei, A.; Wong, P.; Wexler, R. R.; Priestley, E. S.
Atropisomer control in macrocyclic factor VIIa inhibitors. J. Med.
Chem. 2016, 59, 4007−4018.
transition state force field from quantum mechanical calculations. J.
Chem. Theory Comput. 2016, 12, 1833−1844.
(22) Carloni, P.; Alber, F. Quantum medicinal chemistry. In Methods
and Principles in Medicinal Chemistry; Mannhold, R., Kubinyi, H.,
Folkers, G., Eds.; Wiley-VCH: Weinheim, Germany, 2003.
(23) (a) Becke, A. D. Density-functional thermochemistry. III. The
role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (b) Becke,
A. D. A new mixing of Hartree-Fock and local density-functional
theories. J. Chem. Phys. 1993, 98, 1372−1377. (c) Lee, C. T.; Yang, W.
T.; Parr, R. G. Development of the Colle-Salvetti correlation-energy
formula into a functional of the electron-density. Phys. Rev. B: Condens.
Matter Mater. Phys. 1988, 37, 785−789.
(8) LaPlante, S. R.; Fader, L. D.; Fandrick, K. R.; Fandrick, D. R.;
Hucke, O.; Kemper, R.; Miller, S. P. F.; Edwards, P. J. Assessing
atropisomer axial chirality in drug discovery and development. J. Med.
Chem. 2011, 54, 7005−7022.
(9) LaPlante, S. R.; Edwards, P. J.; Fader, L. D.; Jakalian, A.; Hucke,
O. Revealing atropisomer axial chirality in drug discovery.
ChemMedChem 2011, 6, 505−513.
(10) Xing, L.; Devadas, B.; Devraj, R. V.; Selness, S. R.; Shieh, H.;
Walker, J. K.; Mao, M.; Messing, D.; Samas, B.; Yang, J. Z.; Anderson,
G. D.; Webb, E. G.; Monahan, J. B. Discovery and characterization of
atropisomer PH-797804, a p38 MAP kinase inhibitor, as a clinical drug
candidate. ChemMedChem 2012, 7, 273−280.
(11) Sadhu, C.; Masinovsky, B.; Dick, K.; Sowell, C. G.; Staunton, D.
E. Essential role of phosphoinositide 3- kinase delta in neutrophil
directional movement. J. Immunol. 2003, 170, 2647−2654.
(12) Everts, J. B.; Ulrich, R. G. Atropoisomers of 2-Purinyl-3-tolyl-
quinazolinone derivatives and methods of use. WO 2010/111432,
2010.
(13) Marahatta, A.; Bhandary, B.; Lee, Y. C.; Kim, S. R.; Chae, H- J.
Development and validation of a highly sensitive LC-MS/MS method
for quantification of IC87114 in mice plasma, bronchoalveolar lavage
and lung samples: Application to pharmacokinetic study. J. Pharm.
Biomed. Anal. 2014, 89, 197−202.
(14) Thappali, S. R.; Varanasi, K. V.; Veeraraghavan, S.; Vakkalanka,
S. K.; Mukkanti, K. Simultaneous quantitation of IC87114, Roflumilast
and its active metabolite Roflumilast N-oxide in plasma by LC-MS/
MS: application for a pharmacokinetic study. J. Mass Spectrom. 2012,
47, 1612−1619.
(15) Spadoni, G.; Bedini, A.; Lucarini, S.; Mari, M.; Caignard, D. H.;
Boutin, J. A.; Delagrange, P.; Lucini, V.; Scaglione, F.; Lodola, A.;
Zanardi, F.; Pala, D.; Mor, M.; Rivara, S. Highly potent and selective
MT2 melatonin receptor full agonists from conformational analysis of
1-benzyl-2-acylaminomethyl-tetrahydroquinolines. J. Med. Chem. 2015,
58, 7512−7525.
(16) Bedini, A.; Lucarini, S.; Spadoni, G.; Tarzia, G.; Scaglione, F.;
Dugnani, S.; Pannacci, M.; Lucini, V.; Carmi, C.; Pala, D.; Rivara, S.;
Mor, M. Toward the definition of stereochemical requirements for
MT2-selective antagonists and partial agonists by studying 4-phenyl-2-
propionamidotetralin derivatives. J. Med. Chem. 2011, 54, 8362−8372.
(17) Berndt, A.; Miller, S.; Williams, O.; Le, D. D.; Houseman, B. T.;
Pacold, J. I.; Gorrec, F.; Hon, W.-C.; Liu, Y.; Rommel, C.; Gaillard, P.;
(24) Elstner, M. The SCC-DFTB method and its application to
biological systems. Theor. Chem. Acc. 2006, 116, 316−325.
(25) Walker, R. C.; Crowley, M. F.; Case, D. A. The implementation
of a fast and accurate QM/MM potential method in AMBER. J.
Comput. Chem. 2008, 29, 1019−1031.
(26) Rajamani, R.; Naidoo, K. J.; Gao, J. Implementation of an
adaptive umbrella sampling method for the calculation of multidimen-
sional potential of mean force of chemical reactions in solution. J.
Comput. Chem. 2003, 24, 1775−1781.
(27) Patel, L.; Chandrasekhar, J.; Evarts, J.; Forseth, K.; Haran, A. C.;
Ip, C.; Kashishian, A.; Kim, M.; Koditek, D.; Koppenol, S.; Lad, L.;
Lepist, E. I.; McGrath, M. E.; Perreault, S.; Puri, K. D.; Villasenor, A.
̃
G.; Somoza, J. R.; Steiner, B. H.; Therrien, J.; Treiberg, J.; Phillips, G.
Discovery of orally efficacious phosphoinositide 3-kinase δ inhibitors
with improved metabolic stability. J. Med. Chem. 2016, 59, 9228−9242.
(28) Eyring, H. The activated complex in chemical reactions. J. Chem.
Phys. 1935, 3, 107−115.
(29) Barducci, A.; Bussi, G.; Parrinello, M. Well-tempered
metadynamics: a smoothly converging and tunable free-energy
method. Phys. Rev. Lett. 2008, 100, 020603.
(30) Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen, W.
L. Evaluation and reparametrization of the OPLS-AA force field for
proteins via comparison with accurate quantum chemical calculations
on peptides. J. Phys. Chem. B 2001, 105, 6474−6487.
(31) De Vivo, M.; Masetti, M.; Bottegoni, G.; Cavalli, A. Role of
molecular dynamics and related methods in drug discovery. J. Med.
Chem. 2016, 59, 4035−4061.
́
(32) Vymetal, J.; Vondrasek, J. Metadynamics as a tool for mapping
the conformational and free-energy space of peptides - The alanine
dipeptide case study. J. Phys. Chem. B 2010, 114, 5632−5642.
(33) Laio, A.; Parrinello, M. Escaping free-energy minima. Proc. Natl.
Acad. Sci. U. S. A. 2002, 99, 12562−12566.
(34) (a) Laio, A.; Gervasio, F. L. Metadynamics: a method to
simulate rare events and reconstruct the free energy in biophysics,
chemistry and material science. Rep. Prog. Phys. 2008, 71, 126601.
(b) Scalvini, L.; Vacondio, F.; Bassi, M.; Pala, D.; Lodola, A.; Rivara, S.;
Jung, K. M.; Piomelli, D.; Mor, M. Free-energy studies reveal a
possible mechanism for oxidation-dependent inhibition of MGL. Sci.
Rep. 2016, 6, 31046. (c) Russo, S.; Callegari, D.; Incerti, M.; Pala, D.;
Giorgio, C.; Brunetti, J.; Bracci, L.; Vicini, P.; Barocelli, E.; Capoferri,
L.; Rivara, S.; Tognolini, M.; Mor, M.; Lodola, A. Exploiting free-
energy minima to design novel EphA2 protein-protein antagonists:
from simulation to experiment and return. Chem. - Eur. J. 2016, 22,
8048−8052.
Ruckle, T.; Schwarz, M. K.; Shokat, K. M.; Shaw, J. P.; Williams, R. L.
̈
The p110δ crystal structure: mechanisms for selectivity and potency of
new PI(3)K inhibitors. Nat. Chem. Biol. 2010, 6, 117−124.
(18) Avgy-David, H. H.; Senderowitz, H. Toward focusing
conformational ensembles on bioactive conformations: a molecular
mechanics/quantum mechanics study. J. Chem. Inf. Model. 2015, 55,
2154−2167.
(19) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. Ab
Initio calculation of vibrational absorption and circular dichroism
spectra using density functional force fields. J. Phys. Chem. 1994, 98,
11623−11627.
(35) Callegari, D.; Lodola, A.; Pala, D.; Rivara, S.; Mor, M.; Rizzi, A.;
Capelli, A. M. Metadynamics simulations distinguish short- and long-
residence-time inhibitors of cyclin-dependent kinase 8. J. Chem. Inf.
Model. 2017, 57, 159−169.
(20) (a) Warshel, A.; Levitt, M. Theoretical studies of enzymic
reactions: dielectric, electrostatic and steric stabilization of the
carbonium ion in the reaction of lysozyme. J. Mol. Biol. 1976, 103,
227−249. (b) Field, M. J.; Bash, P. A.; Karplus, M. A combined
quantum mechanical and molecular mechanical potential for molecular
dynamics simulations. J. Comput. Chem. 1990, 11, 700−733.
(36) Hohenstein, E. G.; Chill, S. T.; Sherrill, C. D. Assessment of the
performance of the M05-2X and M06-2X exchange-correlation
functionals for noncovalent interactions in biomolecules. J. Chem.
Theory Comput. 2008, 4, 1996−2000.
(37) Ge, Q.; Moir, L. M.; Trian, T.; Niimi, K.; Poniris, M.; Shepherd,
P. R.; Black, J. L.; Oliver, B. G.; Burgess, J. K. The phosphoinositide 3′-
K
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
kinase p110δ modulates contractile protein production and IL-6
release in human airway smooth muscle. J. Cell. Physiol. 2012, 227,
3044−3052.
H. Fused piperidines as a novel class of potent and orally available
transient receptor potential melastatin type 8 (TRPM8) antagonists. J.
Med. Chem. 2012, 55, 1593−1611.
(55) Sherer, E. C.; Lee, C. H.; Shpungin, J.; Cuff, J. F.; Da, C.; Ball,
R.; Bach, R.; Crespo, A.; Gong, X.; Welch, C. J. Systematic approach to
conformational sampling for assigning absolute configuration using
vibrational circular dichroism. J. Med. Chem. 2014, 57, 477−494.
(38) Jaguar, version 7.8; Schrodinger, LLC: New York, NY, 2011.
̈
(39) Maestro, version 9.2; Schrodinger, LLC: New York, NY, 2011.
̈
(40) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A second generation force field for the simulation of
proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 1995,
117, 5179−5197.
(41) Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.;
Simmerling, C. Comparison of multiple Amber force fields and
development of improved protein backbone parameters. Proteins:
Struct., Funct., Genet. 2006, 65, 712−725.
(42) Kumar, S.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A.;
Rosenberg, J. M. The weighted histogram analysis method for free-
energy calculations on biomolecules. I. The method. J. Comput. Chem.
1992, 13, 1011−1021.
(43) Lodola, A.; Capoferri, L.; Rivara, S.; Tarzia, G.; Piomelli, D.;
Mulholland, A.; Mor, M. Quantum mechanics/molecular mechanics
modeling of fatty acid amide hydrolase reactivation distinguishes
substrate from irreversible covalent inhibitors. J. Med. Chem. 2013, 56,
2500−2512.
(44) Desmond Molecular Dynamics System, version 3.0; D. E. Shaw
Research: New York, NY, 2011.
(45) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. Comparison of simple potential functions for
simulating liquid water. J. Chem. Phys. 1983, 79, 926−935.
(46) Krautler, V.; Van Gunsteren, W. F.; Hunenberger, P. H. A fast
SHAKE: Algorithm to solve distance constraint equations for small
molecules in molecular dynamics simulations. J. Comput. Chem. 2001,
22, 501−508.
(47) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: an N −
log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993,
98, 10089−10092.
(48) Tuckerman, M.; Berne, B. J.; Martyna, G. J. Reversible multiple
time scale molecular dynamics. J. Chem. Phys. 1992, 97, 1990−2001.
(49) Bochevarov, A. D.; Harder, E.; Hughes, T. F.; Greenwood, J. R.;
Braden, D. A.; Philipp, D. M.; Rinaldo, D.; Halls, M. D.; Zhang, J.;
Friesner, R. A. Jaguar: a high-performance quantum chemistry software
program with strengths in life and materials sciences. Int. J. Quantum
Chem. 2013, 113, 2110−2142.
(50) Stephens, P. J.; Devlin, F. J.; Pan, J. J. The determination of the
absolute configuration of chiral molecules using vibrational circular
dichroism (VCD) spectroscopy. Chirality 2008, 20, 643−663.
(51) Halgren, T. A. Merck molecular force field. I. Basis, form, scope,
parameterization, and performance of MMFF94. J. Comput. Chem.
1996, 17, 490−519.
(52) SYBYL-X Molecular Modeling Package, version 2.0; Tripos Inc.:
St. Louis, MO.
(53) Frisch, M. J.; Trucks, H. B.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09;
Gaussian, Inc.: Wallingford, CT, 2009.
(54) Tamayo, N. A.; Bo, Y.; Gore, V.; Ma, V.; Nishimura, N.; Tang,
P.; Deng, H.; Klionsky, L.; Lehto, S. G.; Wang, W.; Youngblood, B.;
Chen, J.; Correll, T. L.; Bartberger, M. D.; Gavva, N. R.; Norman, M.
L
DOI: 10.1021/acs.jmedchem.7b00247
J. Med. Chem. XXXX, XXX, XXX−XXX