UDP and UDP-glucose activate the P2Y14 receptor (P2Y14R) to modulate processes related to inflammation, diabetes, and asthma. A computational pipeline suggested alternatives to naphthalene of a previously reported P2Y14R antagonist (3, PPTN) using docking and molecular dynamics simulations on a hP2Y14R homology model based on P2Y12R structures. By reevaluating the binding of 3 to P2Y14R computationally, two alternatives, i.e., alkynyl and triazolyl derivatives, were identified. Improved synthesis of fluorescent antagonist 4 enabled affinity quantification (IC50s, nM) using flow cytometry of P2Y14R-expressing CHO cells. p-F3C-phenyl-triazole 65 (32) was more potent than a corresponding alkyne 11. Thus, additional triazolyl derivatives were prepared, as guided by docking simulations, with nonpolar aryl substituents favored. Although triazoles were less potent than 3 (6), simpler synthesis facilitated further structural optimization. Additionally, relative P2Y14R affinities agreed with predicted binding of alkynyl and triazole analogues. These triazoles, designed through a structure-based approach, can be assessed in disease models.
UDP and UDP-glucose activate the P2Y14 receptor (P2Y14R) to modulate processes related to inflammation, diabetes, and asthma. A computational pipeline suggested alternatives to naphthalene of a previously reported P2Y14R antagonist (3, PPTN) using docking and molecular dynamics simulations on a hP2Y14R homology model based on P2Y12R structures. By reevaluating the binding of 3 to P2Y14R computationally, two alternatives, i.e., alkynyl and triazolyl derivatives, were identified. Improved synthesis of fluorescent antagonist 4 enabled affinity quantification (IC50s, nM) using flow cytometry of P2Y14R-expressing CHO cells. p-F3C-phenyl-triazole 65 (32) was more potent than a corresponding alkyne 11. Thus, additional triazolyl derivatives were prepared, as guided by docking simulations, with nonpolar aryl substituents favored. Although triazoles were less potent than 3 (6), simpler synthesis facilitated further structural optimization. Additionally, relative P2Y14R affinities agreed with predicted binding of alkynyl and triazole analogues. These triazoles, designed through a structure-based approach, can be assessed in disease models.
Extracellular
uridine-5′-diphosphate (1) and
uridine-5′-diphosphoglucose (2, Chart ) activate the P2Y14 receptor (P2Y14R), a G protein-coupled receptor (GPCR)
belonging to the rhodopsin-like δ-branch, to modulate cell functions
related to inflammation, diabetes, asthma, and other diseases.[1,2] This receptor subtype is a member of the P2Y12R-like
subfamily of nucleotide receptors, which inhibit the production of
3′,5′-cyclic adenosine monophosphate (cAMP) through
Gi protein. The P2Y14R promotes hypersensitivity
in microglial cells,[3] the mobility of neutrophils,[4] the release of mediators from mast cells,[5] inflammation in renal intercalated cells,[6] and mixed effects in insulin function.[7,8] Thus, approaches to novel antagonists of nucleotide signaling at
the P2Y14R would be desirable for exploration as novel
therapeutics. The P2Y14R is also present in the CNS, where
it suppresses release of matrix metalloproteinase-9 (MMP-9) and tumor
necrosis factor (TNF) from astrocytes.[9]
Chart 1
Agonist and Antagonist Ligand Probes of the P2Y14Ra
Endogenous agonists 1 and 2 have functional EC50 values
at the hP2Y14R of 160 and 261 nM, respectively.[45] Antagonist 3 is highly potent at
the hP2Y14R (KB from Schild
analysis = 0.434 nM)[14] but suffers from
adverse physicochemical properties (highly hydrophobic/amphipathic)
that make dissolution and purification difficult. Fluorescent antagonist 4 is highly potent at the hP2Y14R (functional Ki = 0.080 nM).[15].Only a limited set of P2Y14R antagonists
are currently
known. Several chemotypes based on naphthoic acid and pyrido[4,3-d]pyrimidine were reported originally in patents to provide
potent P2Y14R antagonists, which however displayed low
oral bioavailability.[10−13] One of those naphthoic acid derivatives, 4-(4-(piperidin-4-yl)-phenyl)-7-(4-(trifluoromethyl)-phenyl)-2-naphthoic
acid (PPTN, 3), was profiled pharmacologically at the
entire family of eight P2YRs and found to display high affinity and
selectivity (IC50 = 0.4 nM at P2Y14R and >10
μM at other P2YR subtypes).[14] We
demonstrated that the piperidine group of 3 was a suitable
site for chemical derivatization and chain extension to prepare high
affinity fluorescent probes of the P2Y14R. This conclusion
was supported by molecular modeling and ligand docking, which showed
the piperidine ring facing outward at the surface of the receptor.[15] One such probe, 4, displayed exceptionally
high affinity and low nonspecific binding when used as a tracer in
a flow cytometric assay of the P2Y14R in whole Chinese
hamster ovary (CHO) cells expressing the receptor.Although
previous approaches to modeling of P2YRs were subject
to high uncertainty, we now have appropriate templates to obtain detailed
docking predictions and structural explanations of previously determined
structure–activity relationship (SAR) within the P2Y12R-like subfamily, e.g., uracil nucleotides binding to the P2Y14R.[16] In the present study, a human
(h) P2Y14R homology model based on recent hP2Y12R X-ray structures[17,18] served as a template to conduct
docking and molecular dynamics (MD) simulations. The immediate goal
was to suggest bioisosteric alternatives to the hydrophobic and unwieldy
naphthalene ring of 3 that would maintain a similar orientation
of the piperidine and 4-(trifluoromethyl)-phenyl substituents when
bound to the receptor and therefore preserve receptor affinity. We
sought to simultaneously reduce the molecular weight and avoid the
high lipophilicity of 3 that contributes to its low solubility
and difficulty of purification.[13]
Results
The macromolecular starting point of our study is a previously
obtained hP2Y14R MD-refined homology model based on the
agonist-bound hP2Y12R X-ray structure (PDB ID: 4PXZ).[16−18] The reference
antagonist structure 3 (PPTN) has been docked into the
model by using an Induced Fit Docking (IFD) protocol (see Methods
section), and the complex has been refined by subjecting it to 10
ns of membrane MD simulations. With respect to the starting agonist-bound
hP2Y14R homology model (Supporting Information, Figure S1A),[16] the
refined structure used in this study (Supporting Information, Figure S1B) featured a larger binding cavity
extending toward the extracellular side. The adaptation of the binding
site to the antagonist structure caused a rearrangement of the position
of transmembrane domain (TM)2 and TM7 with respect to the TM bundle
(Supporting Information, Figure S2A): in
particular, TM7 was pushed outward, whereas the axis of TM2 became
slightly bent toward TM3. Moreover, the extension of the binding site
region toward the extracellular side reoriented the first and second
extracellular loops (ECL1 and ECL2, respectively) away from the TM
bundle (Supporting Information, Figure S2B,C). Notably, the differences described above in the MD refined
agonist- and antagonist-bound hP2Y14R homology models mirror
those experimentally observed for the hP2Y12R X-ray structures.[17,18] Therefore, the final hP2Y14R MD-refined structure (Supporting Information) has been used as a template
for all subsequent docking simulations.Starting from this refined
template, we developed a computational
pipeline comprising three subsequent phases (Figure ): (i) bioisostere search, (ii) compound
screening, and (iii) hit selection. The bioisostere search stage envisaged
the use of docking runs followed by 30 ns of MD simulations to identify
linking groups as suitable replacements of half of the naphthalene
ring while preserving the ligand–receptor interactions observed
for the reference compound. The selection of the linkers has been
mainly guided by the knowledge gained in a previous study that led
to the design of highly potent A3 adenosine receptor (A3AR) agonists.[19] Once a new scaffold
was found, a small library of hypothetical compounds was screened
by means of docking simulations. Pose filtering and hit selection
were based upon ligand–receptor complementarity and optimal
overlap between ligand functional groups and computed protein interaction
sites.
Figure 1
Workflow of the computational pipeline. Selection of naphthalene
bioisosteres was guided by a previous study focused on A3AR agonists. Between the two proposed alternatives, the triazole
analogue resulted as being the most promising according to membrane
MD simulations analysis of the ligand–protein complexes as
compared with the reference compound. Consequently, a small library
of 57 triazole analogues was designed and docked inside the hP2Y14R. Poses were selected by visual inspection. The synthesis
and experimental validation of the compounds were prioritized according
to the overlap between compounds functional groups and protein interaction
sites.
Workflow of the computational pipeline. Selection of naphthalene
bioisosteres was guided by a previous study focused on A3AR agonists. Between the two proposed alternatives, the triazole
analogue resulted as being the most promising according to membrane
MD simulations analysis of the ligand–protein complexes as
compared with the reference compound. Consequently, a small library
of 57 triazole analogues was designed and docked inside the hP2Y14R. Poses were selected by visual inspection. The synthesis
and experimental validation of the compounds were prioritized according
to the overlap between compounds functional groups and protein interaction
sites.In
the first instance, the parent structure 3 was
redocked into the MD-refined model. The corresponding pose (Figure A) showed an overall
root-mean-square deviation (RMSD) value with respect to the final
MD snapshot of 0.42 Å (data not shown). In the predicted binding
mode, the ligand resided in the orthosteric binding site in an orientation
similar to that previously reported for hP2Y14R antagonist
probes:[15] the 2-naphthoic acid carboxylate
bridged Lys772.60 (according to conventional TM numbering[20]), Lys2777.35, and Tyr1023.33, while the piperidine group pointed outward with the nitrogen atom
establishing a H-bond interaction with the backbone of Gly802.63. The docking pose has been subjected to 30 ns of membrane MD simulation,
and the RMSD with respect to the starting structure has been computed
(Figure B, magenta
line); this analysis served as a confirmation of the docking pose
stability as well as a reference to compare newly proposed compounds
containing bioisosteres of the naphthalene. During the MD simulation,
compound 3 exhibited an average RMSD of 3.76 Å that
can be ascribed mainly to motion of the solvent-exposed piperidine
ring. As shown in the trajectory visualization (Supporting Information, Video S1), the naphthalene core with its carboxylate
group is well anchored in the binding site during the 30 ns time frame
by a tight H-bond network established with Lys772.60, Lys2777.35, and Tyr1023.33. However, the distal piperidinenitrogen suddenly moved apart from the pose predicted by docking and
approached the backbone of Ile170 in the second extracellular loop
(EL2) through the interplay of water molecules.
Figure 2
(A) Docking pose of reference
compound 3 (magenta-colored
carbons) at hP2Y14R. (B) RMSD plots for the considered
compounds during 30 ns of membrane MD simulations. (C) Docking pose
of the alkynyl derivative 11 (orange-colored carbons)
at hP2Y14R. (D) Docking pose of the triazolyl derivative 65 (green-colored carbons) at the hP2Y14R. Side
chains of residues important for ligand recognition are reported as
sticks (dark-cyan carbon atoms). Side chains of residues establishing
either van der Waals or hydrophobic contacts with the ligand are rendered
as transparent surface. H-Bonds are pictured as green solid lines,
whereas π–π stacking interactions as cyan dashed
lines with the centroids of the aromatic rings displayed as cyan spheres.
Nonpolar hydrogen atoms are omitted.
(A) Docking pose of reference
compound 3 (magenta-colored
carbons) at hP2Y14R. (B) RMSD plots for the considered
compounds during 30 ns of membrane MD simulations. (C) Docking pose
of the alkynyl derivative 11 (orange-colored carbons)
at hP2Y14R. (D) Docking pose of the triazolyl derivative 65 (green-colored carbons) at the hP2Y14R. Side
chains of residues important for ligand recognition are reported as
sticks (dark-cyan carbon atoms). Side chains of residues establishing
either van der Waals or hydrophobic contacts with the ligand are rendered
as transparent surface. H-Bonds are pictured as green solid lines,
whereas π–π stacking interactions as cyan dashed
lines with the centroids of the aromatic rings displayed as cyan spheres.
Nonpolar hydrogen atoms are omitted.The first proposed alternative structure was an alkyne derivative
(11, Scheme ) containing a p-CF3-phenyl group,
similar to 3. The compound was docked into the MD-refined
hP2Y14R homology model and then subjected to 30 ns of MD
simulations. In the resulting pose (Figure C), the trifluoromethylphenyl group was buried
in a deep hydrophobic pocket, and the carboxylate group interacted
with Lys772.60, Tyr1023.33, and Lys2777.35 as in the docking pose of 3 (Figure A). Although during the MD simulations of
the hP2Y14–11 complex, the ligand atoms
displayed an average RMSD value (3.79 Å) close to the one observed
for the reference compound, the corresponding graph suggested that
compound 11 was less stable than 3 in the
binding pocket (Figure B, orange and magenta lines, respectively). In particular, in the
first 15 ns of simulation, compound 11 experienced an
increasing deviation from the starting pose, leading to a higher root-mean-square
fluctuation (RMSF = 1.52 vs 1.16 Å for compounds 11 and 3, respectively). The trajectory analysis (Supporting
Information, Video S2) revealed that, after
a few ns, the piperidine ring of derivative 11 moved
from the docking pose to establish an H-bond interaction with the
backbone of Ile170EL2 and Gln169EL2 in a way
similar to that observed for the reference compound. On the other
hand, after approximately 12 ns, the side chain of Tyr1023.33 moved to engage the trifluoromethylphenyl group in a T-shaped π–π
stacking interaction. This movement caused a weakening of the H-bond
network around the carboxylate moiety of 11 that increased
the number of water molecules surrounding the group during the trajectory.
Scheme 1
Synthesis of an Alkynyl Derivative 11 as a P2Y14R Antagonist
Reagents and conditions: (a)
CH3OH, SOCl2, 0–23 °C (33%); (b)
1-ethynyl-4-(trifluoromethyl)benzene, CuI, PdCl2(PPh3)2, DMF, NEt3, 0–23 °C (81%);
(c) (CF3SO2)2O, NEt3, CH2Cl2 (98%);
(d) tert-butyl 4-(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)piperidine-1-carboxylate,
Pd(PPh3)4, K2CO3, DMF
(67%); (e) LiOH (aqueous 0.5M), CH3OH reflux, then HCl
(1 M aq), pH 1 (28%).
Synthesis of an Alkynyl Derivative 11 as a P2Y14R Antagonist
Reagents and conditions: (a)
CH3OH, SOCl2, 0–23 °C (33%); (b)
1-ethynyl-4-(trifluoromethyl)benzene, CuI, PdCl2(PPh3)2, DMF, NEt3, 0–23 °C (81%);
(c) (CF3SO2)2O, NEt3, CH2Cl2 (98%);
(d) tert-butyl 4-(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)piperidine-1-carboxylate,
Pd(PPh3)4, K2CO3, DMF
(67%); (e) LiOH (aqueous 0.5M), CH3OH reflux, then HCl
(1 M aq), pH 1 (28%).On the basis of the
analogy observed for the docking poses and
trajectory analyses of compound 11 with respect to the
reference structure 3, we synthesized 11 using both Sonogashira[21] and Suzuki[22] cross-coupling reactions (Scheme ). In particular, a Sonogashira coupling
reaction of a iodo arene 5b with 4-(trifluoromethyl)phenylacetylene
(6) in the presence of tetrakis(triphenylphosphine)palladium
yielded derivative 7. Compound 5b was obtained
by esterification of a commercial precursor 5a. The 3-hydroxy
group of 7 was then converted to the corresponding aryl
triflate 8 using triflic anhydride and pyridine. A Suzuki
cross-coupling reaction between 8 and dioxaborolane derivative 9, prepared by conventional Suzuki–Miyaura reaction[23] in the presence of tetrakis(triphenylphosphine)palladium
catalyst, afforded compound 10. Finally, removal of the N-Boc protecting group followed by hydrolysis of the ester
provided derivative 11. This route would allow a wide
variety of substitutions to be introduced at a late stage in the synthetic
sequence because of the commercial availability of numerous arylacetylene
derivatives.On the basis of the suggestion that the replacement
of the naphthoic
acid core of 3 with an alkynyl group might lead to the
loss of an anchoring hydrogen bond with Tyr1023.33, other
alternatives were considered as well. A similarly versatile synthetic
approach could be used for introducing arylacetylene moieties in the
form of 1,2,3-triazoles by copper-catalyzed [2 + 3] cycloaddition[24] with an azido group present on the core of the
molecule. The corresponding triazole derivative containing a p-CF3-phenyl group (65, Scheme ) was subjected to
the same computational protocol described for compound 11 (docking followed by 30 ns of MD simulations). The resulting docking
pose (Figure D) suggested
a placement of compound 65 in the binding site similar
to that predicted for 3 and 11, which encompassed
a tight H-bond network around the carboxylate group with the piperidine
ring solvent-exposed. On the other side, a slightly higher placement
within the binding site enabled compound 65 to establish
an additional π–π stacking interaction with the
side chain of His1845.36, while the piperidine ring was
anchored by an H-bond with the Ile167EL2 backbone.
Scheme 2
Synthesis of Triazolyl Derivatives as P2Y14R Antagonists
Reagents and conditions: (a)
CH3OH, SOCl2, 0–23 °C (98%); (b)
Boc2O, DMAP, CH2Cl2; (c) PdCl2(dppf), AcOK, DMF, 95 °C (74%); (d) F3CCO2H, 90 °C (97%); (e) H2, Rh/C, 100 psi (98%);
(f) Pd(Ph3P)4, K2CO3,
DME, 85 °C (71%); (g) (CF3CO)2O, NEt3, Et2O; (h) F3CCO2H, CH2Cl2 (70%); (i) (1) Ts-OH, NaNO2, H2O/acetonitrile, (2) NaN3, (83%); (j) CuSO4, sodium ascorbate (1 M aq); (k) KOH (1 M aq). R is
defined in Tables and 2.
Synthesis of Triazolyl Derivatives as P2Y14R Antagonists
Reagents and conditions: (a)
CH3OH, SOCl2, 0–23 °C (98%); (b)
Boc2O, DMAP, CH2Cl2; (c) PdCl2(dppf), AcOK, DMF, 95 °C (74%); (d) F3CCO2H, 90 °C (97%); (e) H2, Rh/C, 100 psi (98%);
(f) Pd(Ph3P)4, K2CO3,
DME, 85 °C (71%); (g) (CF3CO)2O, NEt3, Et2O; (h) F3CCO2H, CH2Cl2 (70%); (i) (1) Ts-OH, NaNO2, H2O/acetonitrile, (2) NaN3, (83%); (j) CuSO4, sodium ascorbate (1 M aq); (k) KOH (1 M aq). R is
defined in Tables and 2.
Table 1
Selection of Terminal Aryl Group Based
on Docking and Molecular Dynamics Simulation of Various Triazole Derivativesc
Percent
inhibition at 3 μM
of binding of fluorescent antagonist 4 (20 nM) in P2Y14R-CHO cells (n = 3). NS, not synthesized.
Structure in Chart .
All ligands were docked to a
10 ns molecular dynamics refined 3-bound P2Y14R homology model.
Table 2
Second Group of Triazolyl P2Y14R Antagonists That Was
Prepared Based on Expanding the SAR
Found in Table
Percent
inhibition at 3 μM
of binding of fluorescent antagonist 4 (20 nM) in P2Y14R-CHO cells
MD simulations of the hP2Y14–65 complex
resulted in lower average RMSD and RMSF values of 3.48 and 0.70 Å,
respectively (Figure B, dark-green line). After a few ns (Supporting Information, Video S3), the ligand moved deeper in the binding
site and was stabilized by persistent π–π stacking
interactions established between the triazole group and Tyr1023.33 and between the trifluoromethylphenyl group and His1845.36. A π–cation interaction between the triazole
ring and the side chain of Arg2536.55 further contributed
to stabilizing the ligand position during the 30 ns time frame. Conversely
to what was observed for compounds 3 and 11, the piperidine ring moved toward Gly802.63. The carboxylate
group of 65 maintained a stable position, and the water
molecules surrounding the carboxylate group of 11 were
not observed during the simulation of the hP2Y14–65 complex.On the basis of these favorable predictions,
using a strategy similar
to Scheme , triazolyl
derivative 65 and its analogues were synthesized starting
from the 3-amino-5-bromobenzoic acid (12) and 4-(4-bromophenyl)piperidin-4-ol
(16, Scheme ). The carboxylic group of 12 was first converted
to the methyl ester 13, and then the amine function was
protected to give Boc-derivative 14. The palladium-catalyzed
condensation of aryl bromide 14 with bis(pinacolato)-diboron
under basic conditions afforded dioxaborolane 15. The
acid-catalyzed dehydration of 16 yielded alkene 17, which was reduced to provide compound 18.
Derivative 19 was obtained by coupling 18 with compound 15 under Suzuki conditions.[22] The conversion of the amino group of 19 to a trifluoroacetamide 20 was accomplished using trifluoroacetic
anhydride in the presence of triethylamine. Removing the N-Boc protecting group of 20 gave the amine 21. Compound 21 was converted into aryl azide 22 from an arenediazonium tosylate that was generated in situ and subsequent
addition of sodium azide.[25] The protected
1,2,3-triazolyl derivatives 23–63 were synthesized via an azide–alkyne Huisgen cycloaddition
(“click reaction”) involving aryl azide 22, various alkynes, Cu(II)sulfate salt, and sodium ascorbate.[26] One-pot hydrolysis of the trifluoroacetamide
and the ester in the presence of KOH yielded compounds 64–104, which were purified by semiprep HPLC and
isolated either as acetate or triethylammonium salts.The synthesis
of fluorescent antagonist 4 as previously
reported[15] suffered from a low yield in
the final click cycloaddition step to link the azide-functionalized
fluorophore and the alkyne-functionalized pharmacophore. Given the
unusually high affinity of 4 and its low nonspecific
character, we explored an alternate synthesis of 4. This
was necessary to provide a sufficient supply of fluorescent probe 4 for use in routine assays, which we needed for the new putative
antagonists 11 and 64–104. A more efficient route consisted of forming an amide as the ultimate
or penultimate step. Thus, the pharmacophore was functionalized with
an extended amine through a preformed triazole linker to provide intermediate 113 (Scheme A). The coupling of AlexaFluor488 fluorophore[27] and pharmacophore was attempted by two methods, either:
(1) condensation of the fluorophore as a 5-carboxylic acid 115 to the ethyl ester-protected derivative 113 of the
pharmacophore followed by ester saponification, or (2) by reaction
of the fluorophore that was activated in situ as a N-succinimidyl ester 116 with the amino derivative 114, having a deprotected carboxylic acid (Scheme B). However, only the second
synthetic route provided compound 4 and at the same time
improved the reaction yield compared to the previous synthetic method.[15] Furthermore, we explored a different fluorescent
antagonist analogue for possible use in screening, e.g., 130 containing a cyanine-5 (Cy5) fluorophore, but this compound was
considerably less potent at the hP2Y14R in comparison to 4 (Supporting Information).
Scheme 3
Reagents and conditions: (A)
(a) TsCl, NEt3, DMAP, CH2Cl2, rt
15 h (88%); (b) LiBr, DMF, rt, 12 h (82%); (c) HBr 48% soln, 80 °C,
20 h (61%); (d) NaN3, H2O, reflux, 12 h (80%);
(e) SOCl2, EtOH, 0 °C to rt (78%); (f) K2CO3, DMF (92%); (g) CuSO4 (15 mol %), sodium
ascorbate (45 mol %), t-BuOH:H2O:CH2Cl2 (51%); (h) LiOH (aqueous 0.5 M), CH3OH reflux, then HCl (1 M aq), pH 1 (21%). (B) (a) (1) TSTU, N,N-diisopropylethylamine, DMF, (2) LiOH,
0.5 M, MeOH:H2O; (b) TSTU, N,N-diisopropylethylamine, DMF; (c) DMF, water, 0 °C.
Reagents and conditions: (A)
(a) TsCl, NEt3, DMAP, CH2Cl2, rt
15 h (88%); (b) LiBr, DMF, rt, 12 h (82%); (c) HBr 48% soln, 80 °C,
20 h (61%); (d) NaN3, H2O, reflux, 12 h (80%);
(e) SOCl2, EtOH, 0 °C to rt (78%); (f) K2CO3, DMF (92%); (g) CuSO4 (15 mol %), sodium
ascorbate (45 mol %), t-BuOH:H2O:CH2Cl2 (51%); (h) LiOH (aqueous 0.5 M), CH3OH reflux, then HCl (1 M aq), pH 1 (21%). (B) (a) (1) TSTU, N,N-diisopropylethylamine, DMF, (2) LiOH,
0.5 M, MeOH:H2O; (b) TSTU, N,N-diisopropylethylamine, DMF; (c) DMF, water, 0 °C.Alkyne derivative 11 was tested in a functional
assay
of antagonism of the agonist-induced inhibition of cAMP production
in the presence of 30 μM forskolin in Chinese hamster ovary
(CHO-K1) cells stably expressing the hP2Y14R (P2Y14R-CHO cells, using an EC80 concentration of agonist 2 of 316 nM).[14] Under these conditions,
the IC50 values for 11 was 5690 ± 1440
(n = 3). Thus, this alkyne derivative was shown to
be an antagonist with considerably less affinity than reference antagonist 3.Although the alkyne 11 was not deemed
a high priority
lead compound because of its moderate potency, the triazoles achieved
higher affinity. Compound 65 and the other newly synthesized
triazole derivatives were assayed in a flow cytometry competition
assay using whole cells (P2Y14R-CHO), and fluorescent antagonist
ligand 4 (20 nM) as a tracer. Figure A shows typical flow cytometry traces in
the presence of representative triazole inhibitors at a single concentration.
The antagonist affinities of the various analogues were compared by
this method, first by screening at a relatively high single concentration
of inhibitor (3 μM) to identify the most potent analogues. The
reference compound 65 inhibited fluorescent labeling
by 92%; thus, it appeared to be a suitable highly potent lead compound
for exploring the SAR in this series. Those IC50 values
of triazole analogues inhibiting the fluorescent labeling by 80% or
greater were determined in full concentration–response curves,
which were sigmoidal, as shown for representative compounds in Figure B.
Figure 3
Biological characterization
of triazole derivatives. Flow cytometric
analysis (A) of the binding of selected triazolyl derivatives in comparison
to reference naphthoic acid derivative 3 at the P2Y14R expressed in CHO cells, as detected through inhibition
of fluorescent cell labeling with 4. Concentration–response
curves for selected compounds (B) displayed a smooth concentration
dependence of the inhibition. The IC50 values are given
in Table . (C) Effects
of P2Y14R antagonist 65 on cAMP levels in
P2Y14R-expressing CHO cells.
Biological characterization
of triazole derivatives. Flow cytometric
analysis (A) of the binding of selected triazolyl derivatives in comparison
to reference naphthoic acid derivative 3 at the P2Y14R expressed in CHO cells, as detected through inhibition
of fluorescent cell labeling with 4. Concentration–response
curves for selected compounds (B) displayed a smooth concentration
dependence of the inhibition. The IC50 values are given
in Table . (C) Effects
of P2Y14R antagonist 65 on cAMP levels in
P2Y14R-expressing CHO cells.
Table 3
Potencies of P2Y14R Antagonists
in Flow Cytometry Assay Determined with Full Concentration–Response
Curves
compd
P2Y14R potency, IC50 (nM)a
3
6.0 ± 0.1
65
31.7 ± 8.0
70
290 ± 88
71
115 ± 15
75
91 ± 4
76
237 ± 90
77
72.4 ± 14.0
78
481 ± 81
79
224 ± 64
82
162 ± 25
84
131 ± 39
90
228 ± 63
Percent inhibition
of binding of
fluorescent antagonist 4 (20 nM) in P2Y14R-CHO
cells (n = 3), over a concentration range of 10–9 to 10–5 M.
Selection of favored terminal aryl groups in the triazole
series,
other than 4-F3C-Ph (65), was based on predictions
arising from the docking of a library of 57 hypothetical triazole
derivatives (Supporting Information, Table S1), which could be easily synthesized by the same route as for 65. The criteria underlying the hit selection are reported
in Table and are based upon ligand–receptor complementarity
and overlap of scaffold functional groups with computed receptor interaction
sites (Supporting Information, Table S2). Initially, selected entries 66–75 were synthesized for testing using the fluorescent hP2Y14R assay. On the basis of the SAR determined for the initial set,
a second group of P2Y14R antagonists 76–104 was prepared (Table ) and evaluated similarly for
the ability to inhibit fluorescent binding at the hP2Y14R.Percent
inhibition at 3 μM
of binding of fluorescent antagonist 4 (20 nM) in P2Y14R-CHO cells (n = 3). NS, not synthesized.Structure in Chart .All ligands were docked to a
10 ns molecular dynamics refined 3-bound P2Y14R homology model.Percent
inhibition at 3 μM
of binding of fluorescent antagonist 4 (20 nM) in P2Y14R-CHO cellsThe
compounds that at 3 μM displayed inhibition of >80% of
the fluorescent ligand (4, 20 nM) binding to the P2Y14R tended to be phenyl derivatives with 4-CF3 (65), 4-Cl (70), 4-Br (71), 4-n-Pr (77), 4-i-Pr (78), 4-t-Bu (79), 4-n-pentyl-O (82), 4-OCF3 (84),
and 3,4-F2 (90) substituents. Also, a 5-bromothien-2-yl
derivative (75) nearly completely inhibited P2Y14R fluorescent labeling. The compounds that displayed inhibition of
60%–80% contained a thien-3-yl (72) or 5-chlorothien-2-yl
group (74) or were phenyl derivatives with 4-C2H5 (66), 4-CH3 (76), 4-t-Bu (79), 3-CF3 (80), 3-OCF3 (85), 4-F (87), 3,5-diF (91), 4-COCH3 (94), and benzofuran-2-yl (101) substituents. Among the
weakest compounds, with <20% inhibition were 4-CH2OH
(67), 2-OCH3 (83), 3-OH (96), and 3-NH2 (97) phenyl analogues
and pyridyl (98), pyrazinyl (99), and thiazolyl
(102) analogues. Both unsubstituted phenyl analogue 64 and the cyclohexyl analogue 103 inhibited
P2Y14R binding in the intermediate range. Thus, recognizable
patterns of SAR were: polar groups, H-bond donor groups, and heteroatom
(especially N) substitution of an aryl ring were not well tolerated,
while nonpolar phenyl substituents and especially para-substitution of the phenyl ring were favored.The IC50 values of the compounds that were examined
in the fluorescent assay with full concentration–response curves
are given in Table . The IC50 values of the reference
naphthoic acid 3 was 6.0 nM, and the values for 10 triazoles
ranged from 31.7 nM (65) to 481 nM (78).
The potencies were in the rank order of: 3 > 65 > 77 > 75 > 71, 82, 84 > 79, 90, 76, 70 > 78.Percent inhibition
of binding of
fluorescent antagonist 4 (20 nM) in P2Y14R-CHO
cells (n = 3), over a concentration range of 10–9 to 10–5 M.The most potently binding triazole derivative 65 was
shown in measurements of cAMP to be an antagonist at the P2Y14R expressed in CHO cells, similar to known and potent antagonist 3 (Figure C). P2Y14R agonist 2 (1 μM) was applied
in the absence or presence of antagonist 3 (1 μM)
or compound 65 (1 μM) to cells stimulated with
30 μM forskolin. As expected, compound 2 significantly
decreased forskolin-induced cAMP levels, but in the presence of either
compound 3 or 65, P2Y14R agonist 2 did not inhibit cAMP accumulation. We also studied the effect
of compound 65 on calcium mobilization induced by several
other P2YRs. 65 (up to 10 μM) was inactive as either
agonist or antagonist at the hP2Y1R and hP2Y6R expressed in 1321N1 astocytoma cells (Supporting Information, Figure S4). Compounds 11, 65, and 74 were separately evaluated by the Psychoactive
Drug Screening Program (PDSP)[28] at various
P2YRs and found to be inactive (10 μM) as agonist or antagonist
at humanP2Y1, P2Y2, P2Y4, and P2Y11Rs (calcium transients) expressed in 1321N1 astrocytoma cells
and protease-activated receptor (PAR)1 expressed in mouse KOLF cells.Off-target activities for selected compounds were measured by the
PDSP.[28] Compounds 3 and 11 each showed only a few off-target interactions at <10
μM. At these sites, the measured Ki values (μM) of 3 were 6.79 (D3 dopamine
receptor) and 2.75 (δ-opioid receptor), and the Ki values (μM) of 11 were 1.46 (σ1 receptor) and 3.60 (σ2 receptor). All other
receptors, channels, and transporters in the standard diverse screen
of the PDSP were not significant (i.e., <50% binding inhibition
at 10 μM). A representative triazole derivative (74) showed no off-target interactions, but the trifluoromethyl analogue
(65) bound weakly (Ki, μm)
at H1 histamine (0.17) and α2A (1.56)
and α2C (1.32) adrenergic receptors. Thus, only a
few off-target interactions were detected in these chemical series.
Discussion
The aim of this project was the synthesis of a library of novel
triazole-based structures as possible antagonists of the P2Y14R having improved physicochemical properties. A triazole moiety was
proposed as an alternative bioisosteric replacement for the naphthoic
acid core of the potent P2Y14R antagonist 3. On the basis of the results previously achieved with the same substitutions
for A3AR agonists,[19] alkyne
derivatives and triazole derivatives were considered for components
of the core to mimic the favorable interactions present in the naphthoic
acid series. By docking to a homology model of the receptor, the envisaged
structures were predicted to occupy the same binding site within the
P2Y14R as 3, maintaining a similar orientation
of the piperidine and 4-(trifluoromethyl)-phenyl substituents within
the ligand binding pocket and preserving the affinity. Docking and
MD simulation have suggested that the triazole scaffold can form additional
interactions that stabilize the ligand within the receptor binding
pocket. A p-CF3-phenyl group bearing triazole
(65) proved to be of higher affinity than the corresponding
alkyne (11); thus, the triazoles were explored in detail.The synthetic route to the triazole series was versatile to allow
the introduction of a wide range of functional groups on a terminal
aryl substituent late in the synthesis. Both Suzuki and click cycloaddition
reactions were applied sequentially to a benzoic ester moiety. The
triazole derivatives were prepared by a late-stage diversification
strategy, introducing the 1,2,3-triazole moiety at the end of the
synthetic sequence by a copper-catalyzed [2 + 3] cycloaddition between
an azide moiety and various arylacetylene derivatives. The choice
of arylacetylene derivatives was initially guided by docking and MD
studies. The overall yield from aniline 12 to protected
azide 22 was 15%, with the last two steps (click reaction
and deprotection) of variable, but usually high yield. One unexpected
result was the hydrolysis of the 3-cyanophenyl group during the last
deprotection step to a 3-carboxyphenyl group (92).These new compounds were assayed in a convenient flow cytometric
fluorescence competition assay with our previously reported antagonist
probe 4 in P2Y14R-CHO cells, confirming the
general docking predictions. Moreover, the expanded SAR exploration
(Table ) provided
greater insight. The terminal aryl group attached to the triazole
required hydrophobic substitution for high affinity, as the unsubstituted
phenyl analogue (64) was weak in P2Y14R binding.
The general preference for substitution of a phenyl ring at this position
was p- > m- ≥ o-, as evidenced with the methoxy analogues (81 > 68 ≥ 83) and the fluoro analogues
(87 > 88 ≈ 89). Similarly,
there was a preference for p- over m- substitution in trifluoromethyl (65 > 80) and trifluoromethoxy (84 > 85) analogues.
Thus, there is a hydrophobic pocket that tolerates considerable steric
bulk, e.g., the 4-tert-butyl analogue (79), in this region of the receptor, as predicted in ligand docking
to the P2Y14R homology model. However, introducing polar
groups to form predicted H-bonding interactions with specific groups
surrounding this aryl ring failed to enhance affinity. This trend
could be explained by the observation that a residue side chain (Ser1875.39) that was predicted in the docking simulation to form
H-bonds with the ligand polar groups was not available for ligand
interaction when analyzed in the dynamic context of the membrane-embedded
solvated receptor (data not shown). The most potent compounds, with
IC50 values ranging from 32 to 131 nM, had substituents:
4-CF3-ϕ 65 > 4-n-propyl-ϕ 77 > 5-Br-thienyl 75 > 4-Br-ϕ 71, 4-n-pentyloxy-ϕ 82, and 4-CF3O-ϕ 84. Substitution
of the 4-CF3 group of 65 with 4-CH3 in 76 considerably lowered the affinity 7-fold (IC50 237 nM).
These affinities were not as potent as reference compound 4 (IC50 6.0 nM) in the same assay. The lead molecule 3 is a highly selective antagonist of the P2Y14R;[14] although selectivity of these antagonists
with respect to P2Y12R and P2Y13R remains to
be determined, several derivatives were shown to be inactive at all
P2Y1-like receptors.The physicochemical properties
of the triazole derivatives remain
to be determined experimentally. However, an online tool for calculating
small-molecule pharmacokinetic and toxicity properties predicted some
advantage; 3 (2.7 μM solubility predicted) would
be 100% bound to human plasma protein and 65 (6.9 μM
solubility predicted) would be 1.5% unbound.[29] The triazole ring benefited from increased polarity and additional
H-bond accepting groups compared to the naphthalene core of 3. The cLogP of 3 is 5.65, which is more hydrophobic
than the optimal range of ∼2–4, while the corresponding
triazole derivative 65 had a cLogP of 4.59. The halogen
substitution of a potent 3,4-difluorophenyl analogue 90 might impede potential oxidation by CYP450 enzymes in the liver.
Conclusion
In conclusion, we have used structural insights to discover a new
scaffold 3-(4-aryl-1H-1,2,3-triazol-1-yl)-biphenyl)
for P2Y14R antagonists. The high affinity among members
of this chemical series of triazole derivatives provides new tools
to aid in our understanding of P2Y14R pharmacology and
potentially could lead to clinically useful drug candidates for inflammatory,
endocrine, and other conditions.
Experimental
Section
Chemical Synthesis
Reagents and Instrumentation
The
proton and carbon
nuclear magnetic resonance spectra were recorded using Bruker 400
MHz, Bruker 500, or Bruker 600 NMR spectrometer. Purification of final
compounds was performed by preparative HPLC (column: Luna 5 μm
C18(2) 100 Å, LC column 250 mm × 4.6 mm). Method A: eluent
0.1% TFA in water–CH3CN from 100:0 to 70:30 in 45
min with a flow rate of 5 mL/min. Method B: eluent 10 mM triethyammonium
acetate buffer–CH3CN from 80:20 to 20:80 in 40 min,
then 10 mM triethyammonium acetate buffer–CH3CN
from 20:80 to 0:100 in 10 min with a flow rate of 5 mL/min. Purities
of all tested compounds were ≥95%, as estimated by analytical
HPLC (column: Zorbax SB-Aq 5 μm analytical column, 150 mm ×
4.6 mm; Agilent Technologies, Inc.). Method: eluent 5 mM triethyammonium
phosphate monobasic solution–CH3CN from 90:10 to
0:100 in 20 min, then triethyammonium phosphate monobasic solution–CH3CN from 0:100 to 90:10 in 5 min with a flow rate of 1 mL/min.
Peaks were detected by UV absorption (254 nm) using a diode array
detector. All derivatives tested for biological activity showed >95%
purity in the HPLC system. Analytical thin-layer chromatography was
carried out on Sigma-Aldrich TLC plates, and compounds were visualized
with UV light at 254 nm. Silica gel flash chromatography was performed
using 230–400 mesh silica gel. Unless noted otherwise, reagents
and solvents were purchased from Sigma-Aldrich (St. Louis, MO). Compound 3 was synthesized as reported.[14] Low-resolution mass spectrometry was performed with a JEOL SX102
spectrometer with 6 kV Xe atoms following desorption from a glycerol
matrix or on an Agilent LC/MS 1100 MSD, with a Waters (Milford, MA)
Atlantis C18 column. High-resolution mass spectroscopic (HRMS) measurements
were performed on a proteomics optimized Q-TOF-2 (Micromass-Waters)
using external calibration with polyalanine. cLogP was calculated
using ChemDraw Professional (PerkinElmer, Boston, MA, v. 15.0).
To a solution
of AlexaFluor 488 115 (4.44 mg, 7.08 μmol) and N,N-diisopropylethylamine (1.34 μL,
7.72 μmol) in dry DMF (400 μL), TSTU (2.42 mg, 7.72 μmol)
was added at 0 °C. The resulting mixture was allowed to warm
up at rt and stirred for 2–3 h. Then, a solution of 114 (4.5 mg, 6.44 μmol) and N,N-diisopropylethylamine (1.30 μL, 7.08 μmol) in dry DMF
(300 μL) was added, and the reaction was stirred overnight at
rt. After removal of the solvent, the residue was purified by preparative
HPLC (method A, Rt = 24.9 min). The product 4 was obtained as an orange solid after lyophilization (0.8
mg, 10%). MS (ESI, m/z) 1212 [M
– H]−. ESI-HRMS calcd m/z for C62H57F3N2O12S2 1212.3462, found 1212.3459 [M –
H]−. HPLC purity 96.1% (Rt = 5.7 min).
Lithium hydroxide (aqueous
0.5M, 70 μL, 25 μmol) was added to a solution of 10 (13 mg, 23 μmol) in methanol (0.2 mL), and the mixture
was heated at reflux for 1.5 h. During this time, 10 was
completely consumed. The mixture was allowed to cool to 23 °C
and acidified with hydrochloric acid (1 M) until pH 1. The acidified
mixture was stirred for additional 2 h before solvents were removed
under reduced pressure. The residue was subjected to column chromatography
(silica gel), eluting with chloroform/methanol/acetic acid 100/10/1
(v/v) mixture. Hydrochloric acid (1 M) was added to fractions containing
the product, and the solvent was removed under reduced pressure to
provide the desired product 11 as a hydrochloride salt
(3.1 mg, 28%). MS (ESI, m/z) 450
[M + H]+. 1H NMR (400 MHz, DMSO-d6): δ (ppm) = 8.24 (s, 1H), 8.11 (t, J = 1.63 Hz, 1H), 8.08 (s, 1H), 7.85 (s, 1H), 7.79 (td, J = 1.51, 7.78 Hz, 1H), 7.73 (td, J = 1.38, 7.53
Hz, 1H), 7.69 (d, J = 8.03 Hz, 2H), 7.41 (t, J = 7.65 Hz, 1H), 7.35 (d, J = 8.28 Hz,
2H), 3.26 (br. s., 2H), 3.10–3.19 (m, 2H), 2.87 (d, J = 12.55 Hz, 3H), 2.42 (dt, J = 2.64,
7.22 Hz, 2H), 2.33 (td, J = 1.79, 3.70 Hz, 1H).
General Procedure A: Click Cycloaddition Reaction
To
a solution of aryl azide (22, 1 equiv) and aryl alkyne
(1.5 equiv) in 2 mL of THF:water (1:1), sodium ascorbate (freshly
prepared 1 M aqueous solution) and CuSO4 (0.5 equiv) were
sequentially added. The resulting reaction was vigorously stirred
for 12 h at rt. The reaction mixture was then concentrated in vacuo
and purified by flash chromatography (hexane:ethyl acetate = 6:4).
Yellow solid 5.7 mg (99%). MS (ESI, m/z) 499.2 [M + H]+. ESI-HRMS
calcd for C26H26F3N4O3 499.1957, found 499.1959 [M + H]+.
General
Procedure B: Deprotection of Piperidine N and Ester
Hydrolysis
To a solution of protected ester (1 equiv) in
2 mL of MeOH:H2O (1:1), KOH (10 equiv) was added, and the
resulting mixture was heated at 50 °C for 12 h. After removing
the solvents under reduced pressure, the mixture was purified by semipreparative
HPLC (method B).
To a solution of hex-5-yn-1-ol 105 (0.84 mL, 7.64
mmol), triethylamine (1.28 mL, 9.17 mmol), and 4-(dimethylamino)pyridine
(20 mg, 0.15 mmol) in CH2Cl2 (25 mL) at 0 °C
was added p-toluenesulfonyl chloride (1.53 g, 8.02
mmol) in three portions. The reaction mixture was brought to rt. and
stirred for 15 h. Aqueous NaOH (1 N, 15 mL) was added, and the mixture
was vigorously stirred for 15 min at rt. The aqueous phase was extracted
with CH2Cl2. The combined organic fractions
were washed with water, followed by brine, and then dried over Na2SO4, filtered, and concentrated in vacuo to give
the title compound (1.68 g, 88%) as a yellowish oil. 1H
NMR (400 MHz, CDCl3): δ (ppm) = 7.69 (2H, J = 8.0, d), 7.26 (d, J = 8.0 Hz, 2H),
3.96 (t, J = 6.0 Hz, 2H), 2.31 (s, 3H), 2.05–2.09
(m, 2H), 1.84 (t, J = 4.0 Hz, 1H), 1.65–1.68
(m, 2H), 1.42–1.48 (m, 2H). 13C NMR (100 MHz, CDCl3): δ 144.8, 133.0, 129.9, 127.0, 83.4, 69.9, 69.0, 27.7,
24.2, 21.6, 17.7.
6-Bromohex-1-yne (107)
LiBr (1.7 g, 19.6
mmol) was added to a stirred solution of 106 (1.64 g,
6.52 mmol) in dry DMF (20 mL). After the exothermic reaction, the
mixture was stirred at room temperature for 24 h. Then water (25 mL)
was added and the aqueous phase extracted with Et2O (3
× 25 mL). The collected organic fractions were dried over Na2SO4, filtered, and concentrated in vacuo. The residue
was purified by flash chromatography using as eluent hexane:ethyl
acetate (5:1) to afford a colorless oil (0.86 g, 82%). 1H NMR (400 MHz, CDCl3): δ (ppm) = 3.59 (t, J = 6.4, 2H), 2.24 (m, 2H), 1.90–1.98 (m, 3H), 1.66–1.70
(m, 2H). 13C NMR (100 MHz, CDCl3): δ (ppm)
= 83.8, 69.2, 45.6, 29.7, 29.4, 19.7.
6-Bromohexan-1-amine Hydrochloride
(109)
6-Aminohexanol 108 (0.5 g,
4.27 mmol) was slowly added
to a stirring 48% HBr solution (5.1 mL) at 0 °C, and the resulting
mixture was stirred at 80 °C for 20 h. The mixture was neutralized
by adding 2N NaOH (20 mL) and extracted with ethyl acatate (3 ×
20 mL). The combined organic fractions were washed with water (50
mL) followed by brine (50 mL) and then dried over Na2SO4, filtered, and concentrated in vacuo. The obtained viscous
oil was dissolved in 4 M HCl solution in dioxane to give a sticky
solid that was washed with Et2O and then filtered to afford
a yellowish solid (0.55 g, 61%). 1H NMR (400 MHz, MeOD):
δ (ppm) = 3.39 (t, J = 4.0 Hz, 2H), 2.86 (t, J = 8.0 Hz, 2H), 1.74–1.82 (m, 2H), 1.59–1.70
(m, 2H), 1.42–1.46 (m, 4H). 13C NMR (100 MHz, MeOD):
δ (ppm) = 44.2, 32.8, 32.2, 27.2, 27.0, 25.2.
6-Azidohexan-1-amine
(110)
To a solution
of 109 (0.55 g, 2.54 mmol) in water (25 mL), NaN3 (0.49 g, 7.69 mmol) was added, and the resulting mixture
was heated at 100 °C for 12 h. After cooling, 37% ammonia solution
was added until a basic pH was reached, and the aqueous phase was
extracted with Et2O (3 × 20 mL). The organic fractions
were collected, dried over Na2SO4, and filtered,
and the solvent removed under vacuum to give a yellow oil (0.29 g,
80%). 1H NMR (400 MHz, CDCl3): δ (ppm)
= 3.18 (t, J = 8.0 Hz, 2H), 2.62 (t, J = 4.0 Hz, 2H), 2.15 (br s, 2H), 1.50–1.54 (m, 2H), 1.35–1.40
(m, 2H), 1.29–1.34 (m, 4H). 13C NMR (100 MHz, CDCl3): δ (ppm) = 51.4, 41.8, 33.1, 28.8, 26.6, 26.4.
To a solution of 3 (0.50
g, 1.57 mmol) in EtOH (50 mL), thionyl chloride (1.37 mL, 18.84 mmol)
was carefully added over 30 min at 0 °C. The reaction was allowed
to warm up to rt and stirred overnight. The resulting mixture was
quenched by adding 5% NaOH solution (25 mL). Then the solvent was
evaporated under vacuum and the aqueous residue extracted with ethyl
acetate (3 × 20 mL). The collected organic fractions were dried
over Na2SO4, filtered, and the solvent removed
under vacuum. The residue was purified by chromatography using as
eluent CH2Cl2/MeOH/NH4OH (7:3:0.3).
The title compound was obtained as a white solid (0.61 g, 78%). MS
(ESI, m/z) 504 [M + H]+. ESI-HRMS calcd for C31H29F3NO2 504.2150, found 504.2150 [M + H]+. 1H NMR (400 MHz, MeOD): δ (ppm) = 8.58 (s, 1H), 8.26 (s, 1H),
7.83–7.90 (m, 4H), 7.76–7.79 (m, 1H), 7.69 (d, J = 8.0 Hz, 2H), 7.33–7.37 (m, 4H), 4.35 (q, J = 8.0 Hz, 2H), 3.15–3.19 (m, 2H), 2.73–2.79
(m, 3H), 1.81–1.84 (m, 2H), 1.65–1.69 (m, 2H), 1.35
(t, J = 4.0 Hz, 3H).
To a solution of 112 (50
mg, 0.09 mmol) in CH2Cl2:t-BuOH:H2O (1:1:1) (2 mL), compound 110 was added, followed
by copper(II) sulfate pentahydrate (15 mol %) and sodium ascorbate
(45 mol %). The reaction mixture was stirred for 24 h at rt. The solvents
were removed under vacuum and the residue rinsed with 37% ammonia
solution (5 mL) and extracted with ethyl acetate (3 × 8 mL).
The collected organic fractions were dried over Na2SO4, filtered, and the solvent removed under reduced pressure.
The residue was purified by chromatography using as eluent a gradient
of CH2Cl2/MeOH/NH4OH (from 9.5:0.5:0.05
to 7:3:0.3). The title product was obtained as a white solid (32 mg,
51%). MS (ESI, m/z) 726 [M + H]+. ESI-HRMS calcd for C43H51F3N5O2 726.3984, found 726.3995 [M + H]+. 1H NMR (400 MHz, CDCl3): δ (ppm) =
8.62 (s, 1H), 8.15 (s, 1H), 7.96–7.99 (m, 2H), 7.73–7.76
m, 2H), 7.66–7.70 (m, 3H), 7.38–7.40 (m, 2H), 7.30–7.33
(m, 2H), 7.21–7.23 (m, 1H), 4.38 (q, J = 8.0
Hz, 2H), 4.23–4.26 (m, 2H), 3.09–3.17 (m, 2H), 2.65–2.71
(m, 6H), 2.39–2.42 (m, 3H), 2.04–2.07 (m, 2H), 1.83–1.88
(m, 6H), 1.58–1.65 (m, 6H), 1.35–1.39 (m, 7H).
To a solution of 113 (32 mg, 44 μmol) in MeOH (1 mL), a solution of 0.5 M LiOH
(1 mL) was added and the resulting mixture was heated at 60 °C
and stirred overnight. After cooling, the solvents were evaporated
and the residue purified by preparative HPLC (Rt = 31.08 min). The product was obtained as a white solid by
freeze-drying (6.5 mg, 21%). MS (ESI, m/z) 698 [M + H]+. ESI-HRMS calcd for C41H47F3N5O2 698.3690, found 698.3682
[M + H]+. 1H NMR (400 MHz, MeOD): δ (ppm)
= 8.49 (s, 1H), 8.27 (s, 1H), 7.89–7.94 (m, 4H), 7.70–7.72
(m, 4H), 7.66–7.70 (m, 3H), 7.38 (d, J = 8.0
Hz, 2H), 7.32 (d, J = 8.0 Hz, 2H), 4.30 (t, J = 8.0 Hz, 2H), 3.09–3.12 (m, 2H), 2.79 (t, J = 8.0 Hz, 2H), 2.66–2.70 (m, 3H), 2.46–2.50
(m, 2H), 2.19–2.25 (m, 2H), 1.81–1.86 (m, 6H), 1.52–1.64
(m, 6H), 1.26–1.33 (m, 4H).
Pharmacological
Assays
Procedures for functional assays
at the P2Y14R reported previously were followed.[14,15]Ki values were calculated from IC50 values using the Cheng–Prusoff equation.[30]
Fluorescent Ligand Binding
Fluorescent
ligand binding
experiments were performed using FCM for all the final triazole derivatives.
An initial experiment involved screening of the compounds at a single
concentration (3 μM) to determine the percentage inhibition,
followed by a full concentration–response curve for the more
potent compounds.For FCM binding experiments in hP2Y14R-CHO cells using 4 as fluorescent tracer,[15] CHO cells expressing hP2Y14Rs were
grown in 12-well plates and used for the assay when the cells were
80% confluent. CHO cells were incubated with a test compound (nonfluorescent
antagonist) for 30 min, followed by continued incubation for 30 min
after the addition of 20 nM of fluorescent antagonist 4. After the incubation, the cells were prepared for FCM analysis
as described below. Briefly, cells were washed 3 times with ice-cold
PBS and detached with 0.2% EDTA, which was neutralized with media
after 5 min incubation at 37 °C. Cell suspensions were transferred
to polystyrene round-bottom BD Falcon tubes (BD, Franklin Lakes, NJ),
and centrifuged twice at 400g for 5 min. Cells were
then suspended in PBS and proceeded to FCM using a FACSCalibur flow
cytometer (BD, Franklin Lakes, NJ), with a 635 nm laser. MFIs were
recorded in FL-1 channel in log mode. All data were analyzed by nonlinear
least-squares analysis, and Ki values
were calculated using Prism (GraphPad Software, San Diego, CA) for
all assays.To quantify the number of receptor-bound ligands,
we used quantitative
fluorescence calibration. To convert measured fluorescence intensity
(MFI) values into molecules of equivalent soluble fluorochrome (MESF)
values, we used Quantum Alexa fluor 488 MESF calibration beads and
QuickCal program v. 2.3 (Bangs Laboratories, Inc., Fishers, IN) according
to the instructions of the manufacturer. The detailed procedure is
explained in the Supporting Information (Figure S3).Compounds 79, 82, 95, and 101 displayed low solubility in DMSO,
and therefore their
5 mM DMSO-solutions were heated to 50 °C for 5 min prior to the
dilution step. To confirm the physicochemical stability of the derivatives, 1H NMR spectra were taken before and after heating, and no
change in the 1H NMR was observed.
Quantification
of cAMP Accumulation
cAMP accumulation
in cells was quantified by competitive enzyme-linked immunoassay.
Cells were plated in 12-well plates 24 h before the assay, and 30
min before the assay the cells were washed with PBS and the medium
was changed to 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic
acid (HEPES)-buffered serum-free DMEM. The assays were performed in
the presence of 10 μM rolipram and were initiated by the addition
of 30 μM forskolin, agonist and/or antagonists.[14] Incubation was for 15 min at 37 °C and were terminated
by washing the cells twice with ice-cold PBS, and the cells were lysed
using 100 μL of cell lysis buffer (Cell Signaling, Danvers,
MA). Then 50 μL of the cell lysate were used for the quantification
of cAMP using a Cyclic AMP XP assay kit (Cell Signaling, Danvers,
MA) following manufacturer’s instructions.
Molecular
Modeling
Homology Model Refinement
Compound 3 was
built using Maestro and prepared with the Ligprep module[31] using the OPLS-2005 force field.[32] The ligand was docked to the previously obtained
UDP-bound MD-refined hP2Y14R homology model[16] by means of the Induced Fit Docking (IFD) procedure
based on Glide search algorithm and SP scoring function.[33] In particular, a Glide grid was positioned on
the centroid of residues located within 5 Å from the previously
identified cavity.[16] The Glide grid was
built using an inner box of 10 Å × 10 Å × 10 Å
and an outer box that extended 20 Å in each direction around
the inner box. The ligand was docked rigidly, and side chains within
5 Å of the ligand were refined. The top ranked docking pose was
then subjected to 10 ns of MD simulations and the final trajectory
snapshot of the receptor used as a template to redock compound 3 and to dock compounds 11 and 65 along with its structural analogues.
Docking of Compounds 3, 11, and 65
Ligands were
built using Maestro and prepared
using the Ligprep module[31] and OPLS_2005
force field.[32] Molecular docking of ligands
to the 3-bound hP2Y14R MD-refined model was
performed by means of the Glide software package[34] from the Schrodinger suite, using the XP scoring function.[35] The binding site was centered on the ligand
barycenter and a Glide grid was computed (inner box side = 10 Å;
outer box side = 30 Å). The top ranked docking poses for each
ligand were subjected to 30 ns of membrane MD simulations.
Molecular
Dynamics
The setup of the MD simulation was
performed as previously described.[16] In
brief, the hP2Y14R homology model was uploaded to the “Orientations
of Proteins in Membranes (OPM)” server[36] and a suggested orientation was provided based on the 2MeSADP-bound
hP2Y12R orientation (PDB: 4PXZ).[18] The model
was then positioned in a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine
(POPC) lipid bilayer (70 Å × 70 Å) generated by a grid-based
method using the VMD[37] Membrane Plugin
tool, and overlapping lipids within 0.6 Å were removed upon combining
the protein and the membrane system. Each protein–membrane
system was then solvated with TIP3P water using the Solvate 1.0 VMD[37] Plugin tool and neutralized by 0.154 M Na+/Cl– counterions.High-performance
computational capabilities of the Biowulf Linux GPU cluster at the
National Institutes of Health, Bethesda, MD (http://biowulf.nih.gov) were exploited to run MD simulations with periodic boundary conditions
using the Nanoscale Molecular Dynamics (NAMD) software package[38] and the CHARMM36 Force Field.[39,40] The ligands were parametrized by analogy using the ParamChem service
(1.0.0) and implementing the CHARMM General Force Field for organic
molecules (3.0.1).[41,42] A 10000-step conjugate gradient
minimization was initially performed to minimize steric clashes. The
protein and ligand atoms were kept fixed during an initial 8 ns equilibration
of the lipid and water molecules. Atom constraints were then removed,
and the entire system was allowed to equilibrate. The temperature
was maintained at 300 K using a Langevin thermostat with a damping
constant of 3 ps–1. The pressure was maintained
at 1 atm using a Berendsen barostat. An integration time step of 1
fs was used, while hydrogen–oxygen bond lengths and hydrogen–hydrogen
angles in water molecules were constrained using the SHAKE algorithm.[43] VMD 1.9[37] was used
for trajectory analysis and movie making. Each structure was simulated
for 30 ns without constraints. RMSD plots for ligand atoms during
the 30 ns trajectories were used to compare relative ligand stability
in the binding pocket during the simulation.
Docking
of Triazole Analogues
A library of 57 triazole
compounds was designed and screened using the 3-refined
P2Y14 model as receptor template. The binding pose of 3 was set as a constraint for the docking, including exposure
of the positively charged piperidine ring toward the solvent and 2-naphthoic
acid carbonyl interactions with Lys772.60 and Lys2777.35. Molecular docking of ligands to the hP2Y14R MD-refined model was performed by means of the Glide package,[34] using the SP scoring function. The ligand binding
site was defined as above-described. Protein interaction sites were
calculated by means of the SiteMap package[44] as implemented in the Schrodinger suite on the apo form of the hP2Y14R MD-refined model by generating at least 15 points per site
with the OPLS_2005 force field.[32] The top
ranked site was used to select the compounds according to optimal
overlap between hydrophobic and hydrogen bond donor/acceptor protein
sites and ligand functional groups evaluated by visual inspection.
Authors: Richard A Friesner; Robert B Murphy; Matthew P Repasky; Leah L Frye; Jeremy R Greenwood; Thomas A Halgren; Paul C Sanschagrin; Daniel T Mainz Journal: J Med Chem Date: 2006-10-19 Impact factor: 7.446
Authors: Joël Robichaud; Jean-François Fournier; Sébastien Gagné; Jacques Yves Gauthier; Martine Hamel; Yongxin Han; Martin Hénault; Stacia Kargman; Jean-François Levesque; Yaël Mamane; Joseph Mancini; Nicolas Morin; Erin Mulrooney; Jin Wu; W Cameron Black Journal: Bioorg Med Chem Lett Date: 2010-12-28 Impact factor: 2.823
Authors: Fahmi Himo; Timothy Lovell; Robert Hilgraf; Vsevolod V Rostovtsev; Louis Noodleman; K Barry Sharpless; Valery V Fokin Journal: J Am Chem Soc Date: 2005-01-12 Impact factor: 15.419
Authors: Juliana I Sesma; Silvia M Kreda; Natacha Steinckwich-Besancon; Hong Dang; Rafael García-Mata; T Kendall Harden; Eduardo R Lazarowski Journal: Am J Physiol Cell Physiol Date: 2012-06-06 Impact factor: 4.249
Authors: Robert B Best; Xiao Zhu; Jihyun Shim; Pedro E M Lopes; Jeetain Mittal; Michael Feig; Alexander D Mackerell Journal: J Chem Theory Comput Date: 2012-07-18 Impact factor: 6.006
Authors: Fatma Mufti; Young-Hwan Jung; Luigino Antonio Giancotti; Jinha Yu; Zhoumou Chen; Ngan B Phung; Kenneth A Jacobson; Daniela Salvemini Journal: ACS Med Chem Lett Date: 2020-04-30 Impact factor: 4.345
Authors: Anna Junker; Christian Renn; Clemens Dobelmann; Vigneshwaran Namasivayam; Shanu Jain; Karolina Losenkova; Heikki Irjala; Sierra Duca; Ramachandran Balasubramanian; Saibal Chakraborty; Frederik Börgel; Herbert Zimmermann; Gennady G Yegutkin; Christa E Müller; Kenneth A Jacobson Journal: J Med Chem Date: 2019-03-21 Impact factor: 7.446
Authors: Thiago Inácio Teixeira do Carmo; Victor Emanuel Miranda Soares; Jonatha Wruck; Fernanda Dos Anjos; Débora Tavares de Resende E Silva; Sarah Franco Vieira de Oliveira Maciel; Margarete Dulce Bagatini Journal: Inflamm Res Date: 2021-04-27 Impact factor: 4.575
Authors: Young-Hwan Jung; Jinha Yu; Zhiwei Wen; Veronica Salmaso; Tadeusz P Karcz; Ngan B Phung; Zhoumou Chen; Sierra Duca; John M Bennett; Steven Dudas; Daniela Salvemini; Zhan-Guo Gao; Donald N Cook; Kenneth A Jacobson Journal: J Med Chem Date: 2020-08-06 Impact factor: 7.446
Authors: Young-Hwan Jung; Veronica Salmaso; Zhiwei Wen; John M Bennett; Ngan B Phung; David I Lieberman; Varun Gopinatth; John C R Randle; Zhoumou Chen; Daniela Salvemini; Tadeusz P Karcz; Donald N Cook; Kenneth A Jacobson Journal: J Med Chem Date: 2021-03-31 Impact factor: 7.446
Authors: Zhiwei Wen; Veronica Salmaso; Young-Hwan Jung; Ngan B Phung; Varun Gopinatth; Qasim Shah; Alexandra T Patterson; John C R Randle; Zhoumou Chen; Daniela Salvemini; David I Lieberman; Gregory S Whitehead; Tadeusz P Karcz; Donald N Cook; Kenneth A Jacobson Journal: J Med Chem Date: 2022-02-03 Impact factor: 7.446
Authors: Paola Oliva; Mirko Scortichini; Clemens Dobelmann; Shanu Jain; Varun Gopinatth; Kiran S Toti; Ngan B Phung; Anna Junker; Kenneth A Jacobson Journal: Bioorg Med Chem Lett Date: 2021-05-26 Impact factor: 2.940
Chart 1
Figure 1
Figure 2
Scheme 1
Scheme 2
Scheme 3
Figure 3