Summary
Background and Purpose The adenosine A3 receptor (A3R) belongs to a family of four adenosine receptor (AR) subtypes which all play distinct roles throughout the body. A3R antagonists have been described as potential treatments for numerous diseases including asthma. Given the similarity between ARs orthosteric binding sites, obtaining highly selective antagonists is a challenging but critical task.
Experimental approach 39 potential A3R, antagonists were screened using agonist-induced inhibition of cAMP. Positive hits were assessed for AR subtype selectivity through cAMP accumulation assays. The antagonist affinity was determined using Schild analysis (pA2 values) and fluorescent ligand binding. Further, a likely binding pose of the most potent antagonist (K18) was determined through molecular dynamic (MD) simulations and consistent calculated binding free energy differences between K18 and congeners, using a homology model of A3R, combined with mutagenesis studies.
Key Results We demonstrate that K18, which contains a 3-(dichlorophenyl)-isoxazole group connected through carbonyloxycarboximidamide fragment with a 1,3-thiazole ring, is a specific A3R (<1 µM) competitive antagonist. Structure-activity relationship investigations revealed that loss of the 3-(dichlorophenyl)-isoxazole group significantly attenuated K18 antagonistic potency. Mutagenic studies supported by MD simulations identified the residues important for binding in the A3R orthosteric site. Finally, we introduce a model that enables estimates of the equilibrium binding affinity for rapidly disassociating compounds from real-time fluorescent ligand-binding studies.
Conclusions and Implications These results demonstrate the pharmacological characterisation of a selective competitive A3R antagonist and the description of its orthosteric binding mode. Our findings may provide new insight for drug discovery.
What is already known
The search for AR subtype specific compounds often leads to ones with multiple subtype binding
What this study adds
This study demonstrates the pharmacological characterisation of a selective competitive A3R antagonist
MD simulations identified the residues important for binding in the A3R orthosteric site
Clinical significance
This study offers insight into A3R antagonists that may provide new opportunities for drug discovery
- Adenosine A3 receptor
- antagonist
- GPCR
- mutagenesis
- Schild analysis
- functional assay
- molecular dynamics
INTRODUCTION
The adenosine A3 receptor (A3R), belongs to a family of four adenosine receptor (AR) subtypes (A1R, A2AR, A2BR and A3R), and is involved in a range of pathologies including cardiovascular, neurological and tumour-related diseases. In particular, mast cell regulation and myocardial preconditioning are key physiological processes regulated by the A3R (Fredholm et al., 2011). Unsurprisingly therefore, A3R is a pharmaceutical target. Interestingly, the A3R has been described as enigmatic, whereby many of the effects attributed to A3Rs are contradictory (Gessi et al., 2008). Despite this, A3R antagonists having been described as potential treatments of asthma, chronic obstructive pulmonary disease (COPD) and glaucoma (Miwatashi et al., 2008, Okamura et al., 2004, Haeusler et al., 2015), to name a few, and continuous research into both agonists and antagonists at the A3R are warranted. While a number of novel potent and selective A3R antagonists have been previously described (Yaziji et al., 2011, Yaziji et al., 2013, Areias et al., 2019), one of the challenges associated with the druggability of the AR family has been the targeting of individual subtypes with sufficient specificity to limit off-target side effects (Chen et al., 2013).
Although all AR members are activated by the endogenous agonist adenosine, the A2AR and A2BR are predominantly Gs-coupled whereas A1R and A3R generally couple to Gi/o. This classical pathway following A3R activation and Gi/o coupling is the inhibition of adenylate cyclase (AC) resulting in a decrease in cAMP levels. Extracellular signal-regulated kinase 1/2 (ERK1/2) activation has also been described as downstream of A3R and is reported to be dependent on βγ-subunits released from pertussis toxin (PTX)-sensitive Gi/o proteins, phosphatidylinostitol-3-kinase (PI3K), the small GTP binding protein Ras, and MAP/ERK kinase (MEK) (Schulte and Fredholm, 2002).
The A2AR is one of the best structurally characterised G protein-coupled receptors (GPCRs), with multiple crystal structures available (Carpenter et al., 2016, Lebon et al., 2011, Lebon et al., 2015, Xu et al., 2011, Liu et al., 2012, Doré et al., 2011, Jaakola et al., 2008, Cheng et al., 2017). Although the remaining AR subtypes have proven more difficult to crystallise with the A3R structure yet to be resolved, there are a number of A1R structures published (Draper-Joyce et al., 2018, Glukhova et al., 2017, Cheng et al., 2017). Thus, in silico screening of vast compound libraries against receptor structures, known as structural-based drug design (SBDD), offers huge potential in the development of highly selective ligands (Carlsson et al., 2010, Katritch et al., 2010, Lenselink et al., 2016, Lagarias et al., 2018). The limited availability of diverse high-resolution structures of the remaining AR subtypes bound to pharmacologically distinct ligands has meant there is a discrepancy between the capability to predict compound binding versus pharmacological behaviour; partial agonism, inverse agonism, biased agonist, antagonism, allosteric modulation etc (Sexton and Christopoulos, 2018). With this in mind, the potential antagonists (K1-K25, K28 and K35) identified in our previously published virtual screening investigation and binding experiments (Lagarias et al., 2018) and some newly identified potential antagonists (K26, K27, K29-K34 and K36-K39) were pharmacologically characterised using cAMP accumulation. We were able to identify a potent and selective A3R antagonist, K18 (O4-{[3-(2,6-dichlorophenyl)-5-methylisoxazol-4-yl]carbonyl}-2-methyl-1,3-thiazole-4-carbohydroximamide) and, using molecular dynamic (MD) simulations combined with site-directed mutagenesis, elude to the potential binding site. Binding free energy calculations of similar in structure analogs of K18 were consistent with the proposed A3R orthosteric binding area. Kinetic binding experiments of K5, K17 and K18 using a bioluminescence resonance energy transfer (BRET) method combined with functional assays led to the identification of important structural features of K18 for binding and activity. Further evaluation of this compound (and structurally related analogues) may afford a novel therapeutic benefit in pathologies such as inflammation and asthma.
RESULTS
Identification of A3R selective antagonists
We initially conducted a blinded screen of 39 compounds (K1-39) to identify selective A3R antagonists some of which have previously been identified to bind A1R, A3R or A2AR (Lagarias et al., 2018). Our screen was carried out using A3R expressing Flp-In™-Chinese hamster ovary (CHO) cells where cAMP accumulation was detected following a combined stimulation of 10 μM forskolin (to allow A3R mediated Gi/o response to be observed), 1 μM tested compound and the predetermined IC80 concentration of NECA (3.16 nM). Compound K1-39 were identified by unblinding (Table 1 and Supplementary Table 1) but are hereinafter referred to as their denoted ‘K’ number. For the purpose of structure-activity relationships studies, the previously uncharacterised compounds (K26, K27, K29-K34 and K36-K39), were assayed both functionally and through radioligand binding (Supplementary Table 1).
Co-stimulation with 10 μM of both forskolin and NECA reduced the cAMP accumulation when compared to 10 μM forskolin alone and this was reversed with the known A3R antagonist MRS 1220 (Table 1 and Supplementary Fig 1). Compounds K1, K10, K11, K17, K18, K20, K23, K25 and K32 were identified as potential antagonists at the A3R through their ability to elevate cAMP accumulation when compared to forskolin and NECA co-stimulation. Of the nine potential A3R antagonists, eight (excluding K11) appeared to be antagonists at the tested concentration of 1 μM (Supplementary Fig. 2 and Supplementary Table 2).
A number of compounds previously documented (K5, K9, K21, K22 and K24; Lagarias et al., 2018) or determined in this study (K26, K27 and K34) to have sub-micromolar binding affinities for A3R showed no activity in our cAMP-based screen (Table 1, Supplementary Table 1). To ensure robustness of our functional screen, full inhibition curves of NECA in the presence or absence of tested compounds (1 μM or 10 μM) were constructed in A3R Flp-In CHO cells (Supplementary Fig. 3, Supplementary Table. 3). In this preliminary data all nine compounds (K5, K9, K11, K21, K22, K24, K26, K27 and K34) appeared to reduce the NECA potency at the highest tested concentration (10 μM) but showed no effect at 1 μM and thus appear to be low potency antagonists at the A3R.
AR subtype selectivity and specificity
The similarity of the different ARs has meant many compounds display reduced selectivity. Using the A3R Flp-In CHO or CHO-K1 cells expressing A1R, A2AR or A2BR incubated with a single high concentration of antagonist (10 μM) and increasing concentrations of NECA identified K10, K17, K18 and K25 as A3R selective antagonists (Fig. 1). K20 and K23 were antagonists at both the A1R and A3R (Fig. 1 and Table 2). K1, K20 and K23 showed weak antagonism at the A2AR and none of the tested antagonist showed any antagonism of the NECA stimulated response at the A2BR. These selectivity findings agree with our previously published radioligand binding data (Lagarias et al., 2018) and are summarised in Table 2.
Characterisation of competitive antagonists at the A3R
All eight A3R antagonists were confirmed to antagonise IB-MECA (Fig. 2 and Table 3) and preliminary data suggests this extends to NECA antagonism (Supplementary Fig. 4 and Supplementary Table 4) in a concentration-dependent manner. Schild analysis characterised K10, K17, K18, K20, K23, K25 and K32 as competitive antagonists at the A3R (Schild slope not significantly different from unity, Fig. 2). Interestingly, the Schild slope deviated from unity for K1 (in competition experiments with NECA, but not IB-MECA) suggesting a more complicated mechanism of antagonism at the A3R. K20 and K23 were also characterised as competitive antagonists at the A1R (Supplementary Fig. 5 and Supplementary Table 5).
When comparing the activity of A3R selective antagonists (K10, K17, K18 and K25), K18 was the most potent, showed A3R specificity and greater A3R binding affinity (Table 2) and we propose it as our lead compound. The original competition binding experiments that identified the panel of antagonist was performed using [3H]HEMADO (Lagarias et al., 2018). To ensure that the different ligands used in our studies was not influencing our characterisation of the compounds, we assessed the ability of K18 to antagonise cAMP inhibition by HEMADO at the A3R and compared its potency to K17 (Supplementary Fig. 6 and Table 6). In this exploratory data K18 again displayed higher potency than K17 at the A3R.
In addition, we wanted to determine if K18 could also antagonise the activity of the A3R when an alternative downstream signalling component was measured; ERK1/2 phosphorylation (Fig. 3). In line with previously reported findings (Graham et al., 2001; Schulte and Fredholm, 2002), agonists at the A3R increased ERK1/2 phosphorylation after 5 minutes, with IB-MECA 10-fold more potent than NECA (Supplementary Fig. 7) and preliminary data suggests this was entirely Gi/o-mediated (pERK1/2 levels were abolished upon addition of PTX). K18 was able to antagonise A3R-mediated phosphorylation of ERK1/2 with the antagonist potency (pA2 values) not significantly different compared to the cAMP-inhibition assay (Fig. 3C).
A3R constitutive activity and inverse agonism
We next determined if any of our competitive antagonist could function as inverse agonists of the A3R. In our hands, the A3R, when expressed in Flp-In™-CHO cells, displays constitutive activity (Supplementary Fig. 8). All eight characterised A3R antagonists showed a concentration-dependent inverse agonism of the A3R when compared to DMSO control (Fig. 2). This was also found to be the case for DPCPX, K20 and K23 at the A1R (Supplementary Fig. 9). Notably, DMSO showed a concentration-dependent elevation in cAMP accumulation above that of forskolin alone.
MD simulation of the binding mode of K18 at A3R
We next sought to investigate the potential binding pose of K18 within the A3R orthosteric site. Building upon our previous studies where we have generated a homology model of the A3R, K18 was docked into the orthosteric site of the A3R using the GoldScore scoring function and the highest scoring pose was inserted in a hydrated POPE bilayer. The complex was subjected to MD simulations in the orthosteric binding site of A3R with Amber14ff for 100 ns and the trajectory analysed for protein-ligand interactions. We identified a potential binding pose of K18 within the established orthosteric A3R binding pocket (Fig. 4). The MD simulations suggest that K18 forms hydrogen bonds, van der Waals (vdW) and π-π interactions inside the orthosteric binding site of A3R (Fig. 4A). More specifically, MD simulations showed that the 3- (dichlorophenyl) group is positioned close to V1695.30, M1775.40, I2496.54 and L2647.34 of the A3R orthosteric binding site forming attractive vdW interactions. The isoxazole ring is engaged in an aromatic π-π stacking interaction with the phenyl group of F1685.29 (Fig. 4A). The thiazole ring is oriented deeper into the receptor favouring interactions with L2466.51, L903.32 and I2687.39. Hydrogen bonding interactions can be formed between: (a) the amino group of the carbonyloxycarboximidamide molecular segment and the amide side chain of N2506.55; (b) the nitrogen or the sulphur atom of the thiazole ring and N2506.55 side chain (Fig. 4A). For structural comparison and insight, we also modelled K5 and K17 binding at the A3R given the structural similarity: K17 and K18 possess one and two chlorine atoms attached to the phenyl ring of the phenyl-isoxazole system, respectively, whereas K5 has none (Fig. 4B and C).
Molecular Mechanics-Poisson Boltzmann Surface Area (MM-PBSA) calculations validate binding pose of K18
We next applied the MM-PBSA method in the MD simulation trajectories of the compounds to calculate their binding free energies (ΔGeff -see methods for derivation (Stamatis et al., 2019)) and evaluated the energetic contributions for their binding (Table 4). The calculated ranking in the binding free energies were in agreement with experimental differences in potencies (K5 < K17 < K18 < MRS 1220). The known A3R antagonist MRS 1220 displayed the lowest ΔGeff value, which is also in agreement with the experimental potencies. The calculations suggested that the major difference between the energetic components of ΔGeff values for K5, K17 and K18 is on the solvation energies (ΔGsolv). The two chlorine atoms make K18 more lipophilic and reduce the energy required to transfer the compound from solution to the binding area, increasing the free energy of binding and activity compared to K17 and K5.
MD simulation of the binding mode of analogs of K18 at A3R
Significantly, when the 4-thiazolyl in K17 was changed to 2-,3- or 4-pyridinyl in compounds K32, K10, K11 we observed antagonistic activity only for compounds K32 and K10 (not K11) (Table 1) since the pyridine nitrogen in compounds K32 and K10 can interact with N2506.55 due to their proximate positions in binding conformation (see Fig. 4B for K17). This interaction is preserved with both the 2-,3-pyridinyl groups but lost when the nitrogen of the pyridinyl group is in the 4-position. In addition, following MD simulations of compounds K26, K27, K29-K34 and K36-K39, we observed that K26 (Supplementary Fig. 10) displayed a similar binding pose to that of K18 (Fig. 4). However, K26 (and K34 for that matter) which containing a o-diphenylcarbonyl (also present in K18) functionally showed weak antagonistic potency (> 1 μM, Supplementary Fig. 2) suggesting a more complex binding mode is present. We also observed that compared to K26 and K34, the p-substitution in compounds K29 and K36-38 was not favourable for binding at all since this led to a loss of the vdW interaction with the hydrophobic area of the A3R towards TM5 and TM6; as was demonstrated in MD simulations for K36 (Supplementary Fig. 10).
Experimental evaluation of the binding mode of K18 at A3R
Of the identified residues predicted to mediate an interaction between K18 and the A3R, the ones which showed (according to the MD simulations) the most frequent and the most important contacts were chosen for investigation and included amino acids L903.32, F1685.29, V1695.30, M1775.40, L2466.51, I2496.54, N2506.55, L2647.34 and I2687.39 (Fig. 4). Site-directed mutagenesis was performed replacing each residue with an alanine in the A3R and expressed in the Flp-In-CHO™ cells lines. Each mutant was then screened for their ability to suppress forskolin-induced cAMP accumulation in response to NECA/IB MECA stimulation in the presence and absence of K18.
Mutation of residues F1685.29, L2466.51, N2506.55 and I2687.39 abolished agonist induced suppression of forskolin-induced cAMP accumulation and were discontinued in this study (Stamatis et al., 2019). Both L90A3.32 and M177A5.40 showed a significantly decreased NECA and IB-MECA potency. L264A7.34 showed a slight decrease in IB-MECA potency whereas the potency of NECA was similar to WT. Whereas the NECA stimulated cAMP inhibition in V169A5.30 or I249A6.54 expressing Flp-In CHOs was comparable to WT, the IB-MECA stimulated cAMP inhibition was enhanced in potency (Table 5). Mutation of V1695.30 to glutamate, the amino acid present in other AR subtypes, enhanced both NECA and IB-MECA potency.
Further, the affinity (pA2) of K18 at the WT and mutant A3R were compared in order to determine the potential antagonist binding site (Fig. 5, Table 5). K18 displayed no difference in affinity at I249A6.54 (compared to WT), whereas M177A5.40 and V169A5.30 were significantly smaller. Interestingly, we found an increase in the affinity for L90A3.32 and L264A7.34 when compared to WT. As would be expected, the K18 affinity at the A3R mutants was not different between agonists, confirming agonist independence (Supplementary Fig. 11). These experimental findings are reflected in our predicted binding pose of K18 at the WT A3R (Fig. 4).
Kinetics of A3R antagonists determined through BRET
NanoBRET techniques have been successfully used to determine the real-time kinetics of ligand binding to GPCRs (Stoddart et al., 2018, Sykes et al., 2019, reviewed in Soave et al., 2019). Here, we investigated the ability of the selective A3R antagonists MRS 1220, K17 or K18 to inhibit specific binding of the fluorescent A3R antagonist CA200645 to Nluc-A3R (Stoddart et al., 2015, Bouzo-Lorenzo et al., 2019). The kinetic parameters for CA200645 at Nluc-A3R were initially determined as Kon (k1) = 2.86 ± 0.89 × 107 M-1, Koff (k2) = 0.4397 ± 0.014 min-1 with a KD of 17.92 ± 4.45 nM. (Supplementary Fig 12). Our MRS 1220 kinetic data was fitted with the original ‘kinetic of competitive binding’ model (Motulsky and Mahan, 1984; built into GraphPad Prism 8.0) with a determined Kon (k3) and Koff (k4) rate of 3.25 ± 0.28 × 108 M-1 min-1 and 0.0248 ± 0.005 min-1, respectively. This gave a residence time (RT) (RT = 1/Koff) of 40.32 min. It was noticed in the analysis for K5, K17 and K18 that the fit in some cases was ambiguous (Regression with Prism 8: “Ambiguous”, 2019) and/or the fitted value of the compound dissociation rate constant was high (k4 > 1 min-1, corresponding to a dissociation t1/2 of < 42 sec). In order to determine the reliability of the fitted k4 value, data were also analysed using an equation that assumes compound dissociation is too rapid for the dissociation rate constant to be determined reliably and the fits to the two equations were compared (“Kinetics of competitive binding, rapid competitor dissociation”, derived in the Appendix I). This model allowed estimate of the equilibrium binding affinity of the compound (Ki) but not the binding kinetics of K5, K17 and K18 (Supplementary Fig. 13 and Table 4). These affinity values were in agreement with those obtained via Schild analysis (except for K5) and were approximately 10-fold higher than those determined through radioligand binding (Table 4). Notably, the order of affinity (K5 < K17 < K18) was consistent.
Assessing the species selectivity of A3R antagonists
There is a lack of selective antagonists for the rat A3R with only a small number reported including the 6-phenyl-1,4-dihydropyridine MRS 1191 and the trazoloquinazoline MRS 1220 which both bind the human A3R (Jacobson et al., 1997). Comparison of residues of the binding area between human and rat A3R show that they differ in residues 1675.28, 1695.30, 1765.37, 2536.58, 2647.34 (Supplementary Fig. 14). The scarcity of rat A3R antagonists prompted us to investigate if our characterised compounds were also potential antagonists at the rat A3R. Using a Nluc-tagged rat A3R expressing HEK 293 cell line, we conducted NanoBRET ligand-binding experiments whereby we determined the ability of our compounds to inhibit specific binding of the fluorescent antagonist AV039 to Nluc-rat A3R. As expected, AV039 was displaced by increasing concentrations of MRS 1220 (pKi 6.788 ±0.1) (Supplementary Fig. 15 and Supplementary Table 7). We found very weak binding of K17, K18, K10 and K32, with no binding detected below the concentration of 10 μM, whereas K1, K20, K23 and K25 were determined as potential rat A3R antagonists (pKi 6.07 ±0.04, 5.71 ±0.03, 5.93 ±0.04 and 6.37 ±0.06, respectively) (Supplementary Fig. 15 and Supplementary Table 7). K25 had a higher binding affinity for the rat A3R when compared to the human A3R (Table 1) (pKi 6.37 ±0.1 and 5.81, respectively).
MD simulations of the rat A3R (performed as described previously for the human A3R) suggested that the presence of M2647.34 most likely hampers K18 binding due to steric hindrance of the dichloro-phenyl group (Supplementary Fig. 16). In contrast, MD simulations of K25 against rat A3R showed the formation of stable complex (Supplementary Fig. 17). Here, the 2-amido group of the thiophene ring is hydrogen-bonded to the amido group of N2506.55. The thiophene ring forms aromatic π-π stacking interaction with F1685.29 and the 5-(p-chlorophenyl) is oriented deep in the binding area making contacts with L903.32, L2466.51 and W2436.48. M2647.34 forces the large lipophilic moiety of the thiophene ring (3-NHCOCH2SPh(CF3)2) of K25 to locate close to TM2 favouring contacts with A692.61, V722.64, and I2687.39 (Supplementary Fig. 17).
Pharmacokinetic assessments of K18
The metabolic in vitro t1/2 (human liver microsomes, 0.1 mg/mL) of K18 (0.1 μM) was determined (0 to 60 minutes) as 24 minutes and the intrinsic clearance (CLint) calculated as 287.2 μl/min/mg (Supplementary Figure 18). This was comparable to the reference compound verapamil and terfenadine (0.1 μM) with t1/2 determined as 35 and 12 minutes and CLint as 200.1 or 581.1 μl/min/mg, respectively. Human plasma stability assessment determined the percentage of K18 (1 μM) remaining after 120 minutes as 90%, with a t1/2 of >120 minutes. This is considerably higher than the reference compound propantheline (1 μM) which was determined to have a half-life of 55 minutes. The t1/2 of K18 (1 μM) in PBS (pH 7.4) over 240 minutes was determined as >240 minutes, with 87% remaining at 240 minutes and was comparable to the reference compound propantheline (1 μM), with a determined t1/2 of >240 minutes.
DISCUSSION
In silico SBDD efforts in ligand discovery, have proven to be highly successful (Meng et al., 2011). However, given the broad and similar orthosteric binding site of ARs, the search for an AR subtype specific compound often leads to compounds active at more than one of the AR subtypes (Kolb et al., 2012). Given that AR subtypes play distinct roles throughout the body, obtaining highly specific receptor antagonists and agonists is crucial. Here, we presented the pharmacological characterisation of eight A3R antagonists identified though virtual screening. Of these eight compounds, K10, K17, K18, K20, K23, K25 and K32 were determined to be competitive. Whereas K20 and K23 were antagonists at both the A1R and A3R, K10, K17, K18, K25 and K32 were A3R selective antagonists. Indeed, we found no functional activity, or indeed binding affinity (< 30 μM), at the other AR subtypes.
K1, K20 and K23 showed weak antagonism of the A2AR (Figure 1, Table 2). We also tested the compounds at the A2BR and determined none of the compound showed A2BR antagonistic potency (Figure 1, Table 2). These selectivity findings were in agreement with our radioligand binding data (presented here (Supplementary Table 1) and in Lagarias et al., 2018 for K1-25, K28 and K35). However, a number of compounds previously determined to have micromolar binding affinity for A3R (K5, K9, K21, K22, K24, K26, K27 and K34), showed no antagonistic potency in our initial functional screen. Further testing confirmed that these compounds were low potency antagonists and, although supporting the previously published radioligand binding data, confirmed the need for functional testing: not all compounds with binding affinity showed high functional potency.
We showed the A3R, when expressed in Flp-In™-CHO cells, displays constitutive activity. Compounds which preferably bind to the inactive (R) state, decreasing the level of constitutive activity (Giraldo et al., 2007) and in the case of a Gi/o -coupled GPCR leading to an elevated cAMP, are referred to as inverse agonists. All eight characterised A3R antagonists and both characterised A1R antagonists (K20 and K23) were found to act as inverse agonists. We also reported an elevation in cAMP accumulation when cells were stimulated with DMSO, which was concentration-dependent. Given that even low concentrations of DMSO has been reported to interfere with important cellular processes (Tunçer et al., 2018), the interpretation of these data should be made with caution. The initial virtual screening described in Lagarias et al., 2018 was carried out using a combination of a ligand-based and structure-based strategy and the experimental structure of A2AR in complex with the antagonist ZM241385 (PDB ID 3EML) (Jaakola et al., 2008), described as A2AR selective antagonist and inverse agonist (Lebon et al., 2011). Our high hit rate for A3R selective antagonist appears counter-intuitive since the ligand-based virtual screening tool Rapid Overlay of Chemical Structures (ROCS) was used to predict structures similar to ZM241385 (Lagarias et al., 2018). Indeed, ZM241385 has little affinity for A3R and 500-to 1000-fold selectivity for A2AR over A1R. However, as has been previously reported, the search for an AR subtype specific compound often leads to compounds active at multiple AR subtypes (Kolb et al., 2012), likely due to their similar binding site.
We hypothesized that the presence of a chloro substituent in the phenyl ring of 3-phenyl-isoxazole favoured A3R affinity and activity, as following 0Cl < 1Cl < 2Cl i.e. K5 < K17 < K18. This theory is supported by both our radioligand binding, NanoBRET ligand-binding and functional data which determine the relative potency and affinity of the three related compounds K5, K17 and K18 as K5 < K17 < K18. The MD simulations showed that these compounds adopted a similar binding mode at the A3R orthosteric binding site, but the free-energy MM-PBSA calculations showed that K18, having two chlorine atoms and an increased lipophilicity, leaves the solution state more efficiently and enters the highly lipophilic binding area.
For the first time, we have demonstrated the utilisation of a new model which expands on the ‘Kinetic of competitive binding’ model (Motulsky and Mahan, 1984; built into Prism) for fitting fast kinetics data obtained from NanoBRET experiments and assumes the unlabelled ligand rapidly equilibrates with the free receptor. Very rapid competitor dissociation can lead to failure of the fit, eliciting either an ambiguous fit (Regression with Prism 8: “Ambiguous”, 2019) or unrealistically large K3 and K4 values. Whereas we were able to successfully fit the MRS 1220 kinetic data with the Motulsky and Mahan model due to its slow dissociation, fitting of K5, K17 and K18 kinetic data with this model often resulted in an ambiguous fit. Our new model, assuming fast compound dissociation, successfully fit the data and allowed the determination of binding affinity. In the cases where the data was able to fit the Motulsky and Mahan model, the dissociation constant was higher (of the order of 1 min-1), indicating rapid dissociation. Although we found nearly a 10-fold differences in determined binding affinity for MRS 1220, K5, K17 and K18 between BRET ligand binding and radioligand binding assays, we demonstrated the order of affinity remains consistent. Indeed, this was seen across all three experimental approached: Schild analysis, NanoBRET ligand-binding assay and radioligand binding.
Our MD simulations describe the potential binding site of K18, our most potent and selective A3R antagonist, within the A3R orthosteric binding area (Fig. 4A). Here, K18 is stabilised through hydrogen bonding interactions between the amino group and thiazole ring of the ligand and the amide side chain of N2506.55. In addition, the dichloro-phenyl ring can be oriented to the unique lipophilic area of A3R including V1695.30, M1775.40, I2496.54 and L2647.34 stabilized in that cleft through attractive vdW interactions; K18 is further stabilized through π-π aromatic stacking interactions between isoxazole ring and the phenyl group of F1685.29 and the thiazole group is oriented deeper into the receptor favouring interactions with L2466.51 and L903.32 and possibly with I2687.39. The determination of critical residues for antagonist binding becomes particularly difficult in the case of competitive antagonists whereby important amino acids are likely overlapping with those for agonist binding. In combination with our mutagenesis data, the final binding pose of K18 appears to be within the orthosteric binding site, involving residues previously described to be involved in binding of A3R compounds (Arruda et al., 2017). We reported no detectable Gi/o response following co-stimulation with forskolin and NECA or IB-MECA for A3R mutants F168A5.29, L246A6.51, N250A6.55 and I268A7.39 (Stamatis et al., 2019). These findings are in line with previous mutagenesis studies investigating residues important for agonist and antagonist binding at the human A3R (Gao et al., 2002, May et al., 2012). L90A3.32, V169A5.30, M177A5.40, I249A6.54 and L264A7.34 A3R all showed a detectable Gi/o response when stimulated with agonists (Stamatis et al., 2019).
Through performing Schild analysis (results of which were used to inform modelling in Lagarias et al., 2019) we were able to experimentally determine the effect of receptor mutation on antagonist affinity for L90A3.32, V169A/E5.30, M177A5.40, I249A6.54 and L264A7.34 A3R. The pA2 value for I249A6.54 A3R was similar to WT, whereas M177A5.40 and V169A5.30 were significantly smaller suggesting these residues appear to be involved in K18 binding. Interestingly we found an increase in the pA2 for L90A3.32 and L264A7.34 when compared to WT, suggesting an enhanced ability of K18 to act as an antagonist. Further evidence was provided by the MM-PBSA calculations which were in agreement, based on the proposed binding model, between the calculated binding free energy by congeners of K18 having one or no chlorine atoms, i.e. compounds K17 and K5, and the known A3R antagonist MRS 1220, and binding affinities and antagonistic potencies. Importantly, substitution of the 1,3-thiazole ring in K17 with either a 2-pyridinyl ring (K32) or a 3-pyridinyl ring (K10) but not a 4-pyridinyl ring (K11) maintained A3R antagonistic potency. Although we have not directly determined the effects of similar pyridinyl ring substitutions on the higher affinity antagonist K18, we suspect there would be no significant increase in the potency of K18 given the small changes we observed for K17.
The human and rat A3R display 72% homology (Salvatore et al., 1993). Antagonists that are A3R-selective across species or at rat A3R alone are useful pharmacological tools to define the role of these receptors. The lack of rat A3R selective antagonists prompted us to investigate if our characterised A3R antagonists were potential antagonists at the rat A3R. We reported no binding of our lead A3R antagonist, K18, at the rat A3R and MD simulations suggest this is due to steric hinderance by M2647.34. This finding suggests that K18 may not only be A3R specific within the human ARs but may also be selective across species. Of the compounds that showed rat A3R binding (K1, K20, K23 and K25), K25 showed the highest binding affinity and represents an interesting candidate for further investigation. MD simulations showed K25 forms a stable complex with rat A3R and we reported a potential binding pose.
In conclusion, through pharmacological characterisation of a number of potential A3R antagonists, this study has determined K18 as a specific (<1 µM) A3R competitive antagonist. Our mutagenic studies, supported by MD simulations, identified the residues important for K18 binding are located within the orthosteric site of the A3R. Importantly, the absence of a chloro substituent, as is the case in K5, led to affinity loss. Further MD simulations have investigated the selectivity profile of K18 and have demonstrated that K18 failed to bind A1R and A2AR due to a more polar area close to TM5, TM6 when compared to A3R (Lagarias et al., 2019). While a number of novel potent and selective A3R antagonists have been previously described (Yaziji et al., 2011, Yaziji et al., 2013, Areias et al., 2019) we present findings of a unique scaffold which can be used as a starting point for detailed structure-activity relationships (SARs) and represents a useful tool that warrants further assessment for its therapeutic potential.
MATERIALS AND METHODS
Cell culture and Transfection
Cell lines were maintained using standard subculturing routines as guided by the European Collection of Cell Culture (ECACC) and checked annually for mycoplasma infection using an EZ-PCR mycoplasma test kit from Biological Industries (Kibbutz Beit-Haemek, Israel). All procedures were performed in a sterile tissue culture hood using aseptic technique and solutions used in the propagation of each cell line were sterile and pre-warmed to 37°C. All cells were maintained at 37°C with 5% CO2, in a humidified atmosphere. This study used CHO cell lines as a model due to the lack of endogenous AR subtype expression (Brown et al., 2008). CHO-K1-A1R, CHO-K1-A2AR, CHO-K1-A2BR and CHO-K1-A3R cells were routinely cultured in Hams F-12 nutrient mix (21765029, Thermo Fisher Scientific) supplemented with 10% Foetal bovine serum (FBS) (F9665, Sigma-Aldrich). Flp-In-CHO cells purchased from Thermo Fisher Scientific (R75807) were maintained in Hams F-12 nutrient mix supplemented with 10% FBS containing 100 μg/mL Zeocin™ Selection Antibiotic (Thermo Fisher Scientific).
Stable Flp-In-CHO cell lines were generated through co-transfection of the pcDNA5/FRT expression vector (Thermo Fisher Scientific) containing the gene of interest and the Flp recombinase expressing plasmid, pOG44 (Thermo Fisher Scientific). Transfection of cells seeded in a T25 flask at a confluency of ≥80% was performed using TransIT®-CHO Transfection Kit (MIR 2174, Mirus Bio), in accordance with the manufacturer’s instructions. Here, a total of 6 μg of DNA (receptor to pOG44 ratio of 1:9) was transfected per flask at a DNA:Mirus reagent ratio of 1:3 (w/v). 48 hours post-transfection, selection using 600 μg/mL hygromycin B (Thermo Fisher Scientific) (concentration determined through preforming a kill curve) was performed for two days prior to transferring the cells into a fresh T25 flask. Stable Flp-In-CHO cell lines expressing the receptor of interest were selected using 600 μg/mL hygromycin B whereby the media was changed every two days. Successful mutant cell line generation for non-signalling mutants were confirmed by Zeocin™ sensitivity (100 μg/mL).
The Nluc-tagged human A3R expressing HEK 293 cell line along with the Nluc-tagged rat A3R pcDNA3.1+ construct for the generation of stable Nluc-tagged rat A3R expressing HEK 293 cells were kindly gifted to us by Stephen Hill and Stephen Briddon (University of Nottingham). HEK 293 cells in a single well of 6-well plate (confluency ≥80%) were transfected with 2 μg of DNA using polyethyleneimine (PEI, 1 mg/ml, MW = 25,000 g/mol) (Polysciences Inc) at a DNA:PEI ratio of 1:6 (w/v). Briefly, DNA and PEI were added to seporate sterile tubes contianing 150 mM sodium chloride (NaCl) (total volume 50 μl), allowed to incubate at room termperature for 5 minutes, mixing together and incubating for a further 10 minutes prior to adding the combined mix dropwise to the cells. 48 hours post-transfection, stable Nluc-rat A3R expressing HEK 293 cell were selected using 600 μg/mL Geneticin (Thermo Fisher Scientific) whereby the media was changed every two days. HEK 293 cell lines were routinely cultured in DMEM/F-12 GlutaMAX™ (Thermo Fisher Scientific) supplemented with 10% FBS (F9665, Sigma-Aldrich).
Constructs
The human A3R originally in pcDNA3.1+ (ADRA3000000, cdna.org) was cloned into the pcDNA5/FRT expression vector and co-transfected with pOG44 to generate a stable Flp-In-CHO cell line. Mutations within the A3R were made using the QuikChange Lightening Site-Directed Mutagenesis Kit (Agilent Technologies) in accordance with the manufacturer’s instructions. The Nluc-tagged rat A3R pcDNA3.1+ construct, used in the generation of the stable Nluc-tagged rat A3R expressing HEK 293 cell line was kindly gifted to us by Stephen Hill and Stephen Briddon (University of Nottingham). All oligonucleotides used for mutagenesis were designed using the online Agilent Genomics ‘QuikChange Primer Design’ tool and are detailed in Stamatis et al., 2019 (Table S4) and purchased from Merck. All constructs were confirmed by in-house Sanger sequencing.
Compounds
Adenosine, NECA ((2S,3S,4R,5R)-5-(6-aminopurin-9-yl)-N-ethyl-3,4-dihydroxyoxolane-2-carboxamide), IB-MECA ((2S,3S,4R,5R)-3,4-dihydroxy-5-[6-[(3-iodophenyl)methylamino]purin-9-yl]-N-methyloxolane-2-carboxamide), HEMADO ((2R,3R,4S,5R)-2-(2-hex-1-ynyl-6-methylaminopurin-9-yl)-5-(hydroxymethyl)oxolane-3,4-diol), DPCPX (8-cyclopentyl-1,3-dipropyl-7H-purine-2,6-dione) and MRS 1220 (N-(9-chloro-2-furan-2-yl-[1,2,4]triazolo[1,5-c]quinazolin-5-yl)-2-phenylacetamide) were purchased from Sigma-Aldrich and dissolved in dimethyl-sulphoxide (DMSO). CA200645, a high affinity AR xanthine amine congener (XAC) derivative containing a polyamide linker connected to the BY630 fluorophore, was purchased from HelloBio (Bristol, UK) and dissolved in DMSO.
AV039, a highly potent and selective fluorescent antagonist of the human A3R based on the 1,2,4-Triazolo[4,3-a]quinoxalin-1-one linked to BY630 (Vernall et al., 2012), was kindly gifted to us by Stephen Hill and Stephen Briddon (University of Nottingham). PMA was purchased from Sigma-Aldrich. Compounds under investigation were purchased from e-molecules and dissolved in DMSO. The concentration of DMSO was maintained to <1.5% across all experiments (1.26% for all cAMP assays, 1% for pERK1/2 assays and 1.02% or 1.1% for NanoBRET ligand-binding experiments using CA200645 or AV039, respectively).
cAMP accumulation assay
For cAMP accumulation (A2AR and A2BR) or inhibition (A1R or A3R) experiments, cells were harvested and re-suspended in stimulation buffer (PBS containing 0.1% BSA and 25 μM rolipram) and seeded at a density of 2,000 cells per well of a white 384-well Optiplate and stimulated for 30 minutes with a range of agonist concentrations. In order to allow the A1R/A3R mediated Gi/o response to be determined, co-stimulation with forskolin, an activator of AC (Zhang et al., 1997), at the indicated concentration (depending on cell line) was performed. For testing of potential antagonists, cells received a co-stimulation stimulated with forskolin, agonist and compound/DMSO control. cAMP levels were then determined using a LANCE® cAMP kit as described previously (Knight et al, 2016). In order to reduce evaporation of small volumes, the plate was sealed with a ThermalSeal® film (EXCEL Scientific) at all stages.
Phospho-ERK assay
ERK1/2 phosphorylation was measured using the homogeneous time resolved fluorescence (HTRF)® Phospho-ERK (T202/Y204) Cellular Assay Kit (Cisbio Bioassays, Codolet, France) two-plate format in accordance with the manufacturer’s instructions. A3R expressing Flp-In-CHO were seeded at a density of 2,000 cells per well of a white 384-well Optiplate and stimulated with agonist and test compounds for 5 minutes at 37°C. Plate reading was conducted using a Mithras LB 940 (Berthold technology). All results were normalised to 5 minutes stimulation with 1 μM PMA, a direct protein kinase C (PKC) activator (Jiang and Fleet, 2012). To determine if the measured pERK1/2 level was Gi-mediated, we treated cells with Pertussis toxin (PTX) (Tocris Biosciences) for 16 hours at 100 ng/mL prior to pERK assay.
Radioligand Binding
All pharmacological methods followed the procedures as described in the literature (Klotz et al., 1998). In brief, membranes for radioligand binding were prepared from CHO cells stably transfected with hAR subtypes in a two-step procedure. In the first step, cell fragments and nuclei were removed at 1000 x g and then the crude membrane fraction was sedimented from the supernatant at 100000 x g. The membrane pellet was resuspended in the buffer used for the respective binding experiments and it was frozen in liquid nitrogen and stored at −80°C. For radioligand binding at the A1R, 1 nM [3H]CCPA was used, for A2AR 10 nM [3H]NECA and for A3R 1 nM [3H]HEMADO. Non-specific binding of [3H]CCPA was determined in the presence of 1 mM theophylline and in the case of [3H]NECA and [3H]HEMADO 100 μM R-PIA was used. Ki values from competition experiments were calculated using Prism (GraphPad Software, La Jolla, CA, U.S.A.) assuming competitive interaction with a single binding site. The curve fitting results (see Fig. 8 in Lagarias et al., 2018) showed R2 values ≥ 0.99 for all compounds and receptors, indicating that the used one-site competition model assuming a Hill slope of n=1 was appropriate.
NanoBRET ligand-binding
Through the use of NanoBRET, real-time quantitative pharmacology of ligand-receptor interactions can be investigated in living cells. CA200645, acts as a fluorescent antagonist at both A1R and A3R with a slow off-rate (Stoddart et al., 2012). Using an N-terminally NanoLuc (Nluc)-tagged A3R expressing cell line, competition binding assays were conducted. The kinetic data was fitted with the ‘kinetic of competitive binding’ model (Motulsky and Mahan, 1984; built into Prism) to determine affinity (pKi) values and the association rate constant (Kon) and dissociation rates (Koff) for unlabelled A3R antagonists. In several cases this model resulted in an ambiguous fit (Regression with Prism 8: “Ambiguous”, 2019). We developed a new model which expands on the ‘kinetic of competitive binding’ model to accommodate very rapid competitor dissociation, assuming the unlabelled ligand rapidly equilibrates with the free receptor. This method allows determination of compound affinity (pKi) from the kinetic data.
In order to identify if the characterised compounds also bound the rat A3R, we conducted competition binding assays using Nluc-tagged rat A3R expressing HEK 293 cells and the fluorescent compound AV039 (Vernall et al., 2012) rather than xanthine based CA200645, which have previously been reported as inactive at rat A3R (Siddiqi et al., 1996). For both human and rat A3R experiments, filtered light emission at 450 nm and > 610 nm (640-685 nm band pass filter) was measured using a Mithras LB 940 and the raw BRET ratio calculated by dividing the 610 nm emission with the 450 nm emission. Here, Nluc on the N-terminus of A3R acted as the BRET donor (luciferase oxidizing its substrate) and CA200645/AV039 acted as the fluorescent acceptor. CA200645 was used at 25 nM, as previously reported (Stoddart et al., 2015) and AV039 was used at 100 nM (pre-determined KD, 102 ± 7.59 nM). BRET was measured following the addition of the Nluc substrate, furimazine (0.1 μM). Nonspecific binding was determined using a high concentration of unlabelled antagonist, MRS 1220 at 10 nM or 10 μM, for human and rat A3R, respectively.
Receptor binding kinetics data analysis
Specific binding of tracer vs time data was analysed using the Motulsky and Mahan method (Motulsky and Mahan, 1984; built into Prism) to determine the test compound association rate constant and dissociation rate constant. Data were fit to the “Kinetics of competitive binding” equation in Prism 8.0 (GraphPad Software Inc, San Diego, CA): where,
[RL]t is specific binding at time t, N the Bmax, [L] the tracer concentration, [I] the unlabelled competitor compound concentration, k1 the tracer association rate constant, k2 the tracer dissociation rate constant, k3 the compound association rate constant and k4 the compound dissociation rate constant.
Data were also analysed using an equation that assumes compound dissociation is too rapid for the dissociation rate constant to be determined reliably and the fits to the two equations compared (“Kinetics of competitive binding, rapid competitor dissociation”, derived in the Appendix I, Supplementary material). This equation assumes rapid equilibration between compound and receptor and consequently provides an estimate of the equilibrium binding affinity of the compound (Ki) but not the binding kinetics of the compound. The equation is, where ρI is fractional occupancy of receptors not bound by L: and kobs,+ I is the observed association rate of tracer in the presence of competitor, defined as,
The fits to the two equations were compared statistically using a partial F-test in Prism 8 (Motulsky, 2019).
Pharmacokinetic assessments of K18
Preliminary pharmacokinetic assessments of K18 was out-sourced to Eurofins Panlabs (Missouri, U.S.A) and including tests for intrinsic clearance (human liver microsomes), plasma (human) stability and half-life in PBS. These tests were conducted in duplicate using a single concentration of K18 (0.1 μM or 1 μM) using the substrate depletion method. Here, the percentage of K18 remaining at various incubation times was detected using high-performance liquid chromatography mass spectrometry (HPLC-MS). Reference compounds (verapamil, terfenadine and propantheline) were supplied and tested alongside K18. The half-life (t1/2) was estimated from the slope (k) of percentage compound remaining (In(%K18 remaining)) versus time (t1/2 = - In(2)/k), assuming first order kinetics. The intrinsic clearance (CLint, in μl/min/mg) was calculated according to the following formula:
Data and Statistical analysis
All in vitro assay data was analysed using Prism 8.0 (GraphPad software, San Diego, CA), with all dose-inhibition or response curves being fitted using a 3-parameter logistic equation to calculate response range or Emax and IC/EC50. Experimental design ensured random distribution of treatments across 96/384-well plates to avoid systematic bias. Agonist stimulation alone was used as an intrinsic control across all experiments. Although initial screening of the 50 compounds was blinded, due to limitations in resources, it was not possible to perform all of our experiments in a blinded manner. Normalisation was used to control for unwanted sources of variation between experimental repeats.
In the rare cases where significant outliers were identified through the ROUT method (performed in Prism with Q set to 2% (defines the chance of falsely identifying one or more outliers)) (Statistics with Prism 7: “How to: Identify outliers”, 2019), these were excluded from data analysis and presentation. Dose-inhibition/dose-response curves were normalised to forskolin, expressed as percentage forskolin inhibition for Gi-coupled A1R and A3R (1 μM or 10 μM, respectively) or stimulation for A2AR and A2BR (100 μM, representing the maximum cAMP accumulation of the system), relative to NECA/IB-MECA (agonist allowing comparison across AR subtypes and a selective A3R agonist, respectively). For cAMP experiments on A3R mutants, data was normalised to 100 μM forskolin, representing the maximum cAMP accumulation possible for each cell line. In the case of pERK1/2 response, normalisation was performed to PMA, a direct PKC activator providing the maximum pERK1/2 level of the system.
Schild analysis was performed to obtain pA2 values (the negative logarithm to base 10 of the molar concentration of an antagonist that makes it necessary to double the concentration of the agonist to elicit the original submaximal response obtained by agonist alone (Schild, 1947)) for the potential antagonists. In cases where the Schild slope did not differ significantly from unity, the slope was constrained to unity giving an estimate of antagonist affinity (pKB). pA2 and pKB coincide when the slope is exactly unity. Through performing Schild analysis, whereby the pA2 is independent of agonist, we were able to experimentally determine the effect of receptor mutation on antagonist binding. The pA2 values obtained through conducting Schild analysis of K18 at WT and mutant A3R were compared in order to indicate important residues involved in K18 binding. Whereas an increase in the pA2 for a particular mutant when compared to WT suggested the antagonist was more potent, a decrease indicated a reduced potency.
All experiments were conducted in duplicate (technical replicates) to ensure the reliability of single values. The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2018). Statistical analysis, performed using Prism 8.0, was undertaken for experiments where the group size was at least n = 5 and these independent values used to calculate statistical significance (*, p< 0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001) using a one-way ANOVA with a Dunnett’s post-test for multiple comparisons or Students’ t-test, as appropriate. Any experiments conducted n < 5 should be considered preliminary. Compounds taken forward for further investigation after our preliminary screening (n = 3) were selected based on a high mean cAMP accumulation (>80%).
Computational biochemistry
MD simulations
Preparation of the complexes between human A3R with K5, K17, K18 or MRS 1220 and rat A3R with K18 or K25 was based on a homology model of A2AR (see Appendix II in Supplementary material). Each ligand-protein complex was embedded in hydrated POPE bilayers. A simulation box of the protein-ligand complexes in POPE lipids, water and ions was built using the System Builder utility of Desmond (Desmond Molecular Dynamics System, version 3.0; D.E. Shaw Res. New York, 2011; Maest. Interoperability Tools, 3.1; Schrodinger Res. New York, 2012.). A buffered orthorhombic system in 10 Å distance from the solute atoms with periodic boundary conditions was constructed for all the complexes. The MD simulations were performed with Amber14 and each complex-bilayer system was processed by the LEaP module in AmberTools14 under the AMBER14 software package (Case et al., 2014). Amber ff14SB force field parameters (Maier et al., 2015) were applied to the protein, lipid14 to the lipids (Dickson et al., 2014), GAFF to the ligands (Wang et al., 2004) and TIP3P (Jorgensen et al., 1983) to the water molecules for the calculation of bonded, vdW parameters and electrostatic interactions. Atomic charges were computed according to the RESP procedure (Bayly et al., 1993) using Gaussian03 (Frisch et al., 2003) and antechamber of AmberTools14 (Case et al., 2014). The temperature of 310 K was used in MD simulations in order to ensure that the membrane state is above the main phase transition temperature of 298 K for POPE bilayers (Koynova and Caffrey, 1998). In the production phase, the relaxed systems were simulated in the NPT ensemble conditions for 100 ns. The visualization of produced trajectories and structures was performed using the programs Chimera (Pettersen et al., 2004) and VMD (Humphrey et al., 1996). All the MD simulations were run on GTX 1060 GPUs in lab workstations or on the ARIS Supercomputer.
MM-PBSA calculations
Relative binding free energies of the complexes between K5, K17, K18, MRS 1220 and A3R was estimated by the 1-trajectory MM-PBSA approach (Massova and Kollman, 2000). Effective binding energies (ΔGeff) were computed considering the gas phase energy and solvation free energy contributions to binding. For this, structural ensembles were extracted in intervals of 50 ps from the last 50 ns of the production simulations for each complex. Prior to the calculations all water molecules, ions, and lipids were removed, and the structures were positioned such that the geometric centre of each complex was located at the coordinate origin. The polar part of the solvation free energy was determined by calculations using Poisson-Boltzmann (PB) calculations (Homeyer and Gohike, 2013). In these calculations, a dielectric constant of εsolute = 1 was assigned to the binding area and εsolute = 80 for water. Using an implicit solvent representation for the calculation of the effective binding energy is an approximation to reduce the computational cost of the calculations. The binding free energy for each complex was calculated using equation (1)
In equation (1) ΔGeff is the binding free energy for each calculated complex neglecting the effect of entropic contributions or assuming to be similar for the complexes studied. ΔEMM defines the interaction energy between the complex, the protein and the ligand as calculated by molecular mechanics in the gas phase. ΔGsol is the desolvation free energy for transferring the ligand from water in the binding area calculated using the PBSA model. The terms for each complex ΔEMM and ΔGsol are calculated using equations (2) and (3)
In equation (2) ΔEelec and ΔEvdW are the electrostatic and the vdW interaction energies, respectively. In equation (3) ΔGP is the electrostatic or polar contribution to free energy of solvation and the term ΔGNP is the non-polar or hydrophobic contribution to solvation free energy. Molecular mechanics energies and the non-polar contribution to the solvation free energy were calculated with the mmpbsa.pl module (Miller et al., 2012) of Amber14 (Case et
Nomenclature of Targets and Ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017).
Conflict of Interest
None for any author
Regression with Prism 8: ‘Ambiguous’. (2019).
https://www.graphpad.com/guides/prism/8/curve-fitting/reg_analysischeck_nonlin_ambiguous.htm
Statistics with Prism 7: “How to: Identify outliers”. (2019).
SUPPORTING INFORMATION
Acknowledgments
We gratefully acknowledge the support of the Leverhulme Trust (RPG-2017-255) (KB and GL) and the BBSRC (BB/M00015X/2) (GL). This research represents part of the Ph.D work of P.L. We thank Chiesi Hellas which supported this research (SARG No 10354) and the State Scholarships Foundation (IKY) for providing a Ph.D fellowship to P.L. (MIS 5000432, NSRF 2014-2020). The work of E.V. is implemented through IKY scholarships programme and co-financed by the European Union (European Social Fund - ESF) and Greek national funds through the action entitled “Reinforcement of Postdoctoral Researchers”, in the framework of the Operational Program “Human Resources Development Program, Education and Lifelong Learning” of the National Strategic Reference Framework (NSRF) 2014 – 2020. This work was supported by computational time granted from the Greek Research & Technology Network (GRNET) in the National HPC facility - ARIS - under project IDs pr005010. We would like to thank Stephen Hill, Stephen Briddon and Mark Soave (University of Nottingham) for gifting the Nluc-A3R cell line, Nluc-rat A3R construct, the fluorescent antagonist AV039 and their technical advice. We are also grateful to Sonja Kachler for her technical assistance.
Appendix I: Adapting the kinetics of competitive binding equation for rapidly-dissociating unlabelled ligands
The Motulsky and Mahan equation is used to measure the association and dissociation rate constant of unlabelled compounds, in competition with labelled ligand for binding to a receptor (Motulsky and Mahan, 1984). When the unlabelled competitor dissociates rapidly, relative to the early time points of the assay, the fitted parameters can indicate a failure of the fit to provide realistic or sufficiently precise estimates of the model parameters. Here an equation is derived for rapidly-dissociating compounds that provides an estimate of the equilibrium binding affinity rather than the binding rate constants of the unlabelled competitor. It is assumed the competitor is at equilibrium throughout the time course of the binding assay. This is represented by the following scheme, where Ki is the equilibrium dissociation constant of the competitor:
R is receptor, L is labelled ligand, I is unlabelled competitor, k1 is the association rate constant and k2 the dissociation rate constant of the labelled ligand. The assumptions underlying the original kinetics of competitive binding equation apply (Motulsky and Mahan, 1984): Only a small fraction (< 10%) of total tracer and compound concentration is bound by receptor (“Zone A”); receptor is exposed simultaneously to both ligands; receptor-ligand binding conforms to a single-site, single-step non-cooperative mass-action mechanism; and unlabelled ligand competitively inhibits tracer binding.
The goal is an equation that describes binding of labelled ligand to receptor over time in the presence of unlabelled competitor. We start with the differential equation for [RL]:
Next, R is substituted by employing the conservation of mass equation: where N is the total concentration of receptor. Solving for [R] gives,
This expression is now substituted into the differential equation for [RL]:
We now substitute [RI] with an expression in terms of [RL], as follows: Since I is at equilibrium with the receptor, the ratio of R to RI is constant, and can be represented by the following expression: where the constant ρI is fractional occupancy by I of accessible receptors, i.e. those not bound by labelled ligand. Next, this equation is rearranged as follows:
The bracketed term on the right-hand side, [R] + [RI] can be re-written in terms of [RL] and the constant N, using the conservation of mass equation for the receptor:
[R] + [RI] is now substituted with N − [RL]:
This is the desired expression: [RI] is expressed in terms of [RL]. This expression is now substituted into the differential equation for [RL], which gives, after rearranging,
This equation is now integrated to obtain the [RL] vs t equation that can be used to fit experimental data: where and
Appendix II: Supporting information for computational biochemistry
Preparation of the structures
Structures of the compounds K5, K17, K18 or MRS 1220 were prepared using Maestro (Version 10.5; Schrodinger, Inc.: New York, NY, 2015) and minimized as in a previous paper (Lagarias et al., 2018). The inactive state homologue of A3R WT was taken from Adenosiland web-service (Floris et al., 2012). The BLAST algorithm estimated the human A1R (Glukhova et al., 2017; PDB ID 5UEN) as the most appropriate template for human A3R model having the most similar sequence. The rat A3R was generated using the protein structure homology model server SwissModel (Waterhouse et al., 2018). We also applied MODELLER 9.18 (Šali and T. L. Blundell, 1993, Eswar et al., 2003) which selected ten PDB structures with the highest sequential similarity as templates for homology modeling of rat A3R. Twenty homology models were generated and the model with the lowest DOPE (Discrete Optimized Protein Energy) value was selected. The two resulting rat A3R models created by Swiss Model and MODELLER 9.18 were compared to each other using the Protein Structure Alignment Tool of Desmond Maestro 2018-1 and were found to be very similar (Desmond Molecular Dynamics System, version 3.0; D.E. Shaw Res. New York, 2011; Maest. Interoperability Tools, 3.1; Schrodinger Res. New York, 2012).
The protein models were optimized as previously published using the Protein Preparation Wizard implementation in Schrodinger suite (Protein Prep. Wizard 2015-2; Epik version 2.4, Schrödinger, LLC, New York, NY, 2015; Impact version 5.9, Schrödinger, LLC, New York, NY, 2015; Prime version 3.2, Schrödinger, LLC, New York, NY, 2015). The ZM241385-inactive A2AR protein complex from 3EML (Jaakola et al., 2008) was superimposed to human or rat A3R WT and the A2AR protein was removed resulting in ZM241385-A3R complex which was used as a template for docking of K5, K17, K18, K25 or MRS 1220 using GoldScore and ChemPLP scoring functions (Jones et al., 1997, Eldridge et al., 1997) and the GOLD (Version 5.2, Cambridge Crystallogr. Data Cent. Cambridge, U.K., 2015) (Verdonk et al., 2005) as previously described (Lagarias et al., 2018). The top high-scoring poses for K5, K17, K18, K25 or MRS 1220 in complex with A3R using GoldScore were better and were kept. These complexes were embedded in POPE bilayers using the System Builder utility of Desmond (Desmond Molecular Dynamics System, version 3.0; D.E. Shaw Res. New York, 2011; Maest. Interoperability Tools, 3.1; Schrodinger Res. New York, 2012). Complex and lipid systems were solvated using the TIP3P water model (Jorgensen, 1983). Na+ and Cl- ions were placed in the water phase to neutralize the systems and to reach the experimental salt concentration of 0.150 M NaCl. A 10 Å-from the solute atoms-buffered orthorhombic system with periodic boundary conditions was constructed for all complexes.
MD simulations
Each ligand-A3R complex in the bilayer was processed by the LEaP module in AmberTools14 under the AMBER14 software package (Case et al., 2014). Amber ff14SB force field parameters (Maier et al., 2015) was applied to the protein, lipid14 to the lipids (Dickson et al., 2014), GAFF to the ligands (Wang et al., 2004) and TIP3P (Jorgensen, 1983) to the water molecules for the calculation of bonded, vdW parameters and electrostatic interactions. Atomic charges were computed according to the RESP procedure (Bayly et al., 1993) using Gaussian03 (Frisch et al., 2003) and antechamber of AmberTools14. MD simulations in explicit solvent were performed using PMEMD (Case et al., 2014). MD simulation protocol consists of five stages: a) Minimization, b) Heating, c) Adjustment of density, d) Equilibration and e) Production. The systems were minimized by 2500 steps of steepest descent to remove bad contacts and 7500 steps of conjugated gradient minimization in the presence of a harmonic restraint with a force constant of 5 kcal mol-1 Å-2 on all atoms of protein and ligand and non-bonded cut-off of 8.0 Å. The next stage in MD simulation protocol is to allow the system to heat up from 0 K to 310 K. Langevin thermostat (dynamics) (Izaguirre et al, 2001) as implemented in Amber14 (Case et al., 2014) was used for temperature control employing a Langevin collision frequency of 2.0 ps-1. The system in two consecutive steps to 310 K in the presence of a harmonic restraint with a force constant of 10 kcal mol-1 Å-2 on all membrane, protein, and ligand atoms. In the first step, systems were heated to 100 K in a NVT of 50 ps length where the adjustment of the density was realized using the Berendsen barostat (Berendsen et al., 1984) with a 2 ps coupling time. In the second step, the temperature was raised to 310 K in a NPTγ (with γ = 10 dyn cm-1) simulation of 500 ps length. Subsequently, the systems were equilibrated without restraints in a NPTγ simulation of 1 ns length with T = 310 K and γ = 10 dyn cm-1. The equilibration phase was followed by production simulation for 100 ns with system-specific lengths using the same protocol as in the final equilibration step. In the NPTγ simulations semiisotropic pressure scaling to p = 1 bar was applied using a pressure relaxation time of 1.0 ps. For the treatment of long-range electrostatic interactions, the Particle-mesh Ewald summation method (Darden et al., 1993, Essmann et al., 1995) was used, and short-range non-bonding interactions were truncated with an 8 Å cutoff. Bonds involving hydrogen atoms were constrained by the SHAKE algorithm (Ryckaert et al., 1977), and a time step of 2 fs was used for the integration of the equations of motion. Snapshots recorded every 20 ps during the production Properties and dynamics of the protein and ligand systems as well as of the membrane were analyzed with the ptraj and cpptraj modules of AmberTools12 (Case et al., 2014).
Abbreviations
- AR
- adenosine receptor
- A1R
- A1 adenosine receptor
- A2AR
- A2A adenosine receptor
- A2BR
- A2B adenosine receptor
- A3R
- A3 adenosine receptor
- CA200645
- fluorescent xanthine amine congener
- cAMP
- adenosine 3’,5’ cyclic monophosphate
- CHO
- Chinese hamster ovary,
- DMSO
- dimethyl sulfoxide
- DPCPX
- 8-cyclopentyl-1,3-dipropyl-7H-purine-2,6-dione
- ERK
- extracellular signal-regulated kinase
- IB-MECA
- (2S,3S,4R,5R)-3,4-dihydroxy-5-[6-[(3-iodophenyl)methylamino]purin-9-yl]-N-methyloxolane-2-carboxamide
- HEMADO
- (2R,3R,4S,5R)-2-(2-hex-1-ynyl-6-methylaminopurin-9-yl)-5-(hydroxymethyl)oxolane-3,4-diol
- MRS 1220
- N-(9-chloro-2-furan-2-yl-[1,2,4]triazolo[1,5-c]quinazolin-5-yl)-2-phenylacetamide
- NECA
- (2S,3S,4R,5R)-5-(6-aminopurin-9-yl)-N-ethyl-3,4-dihydroxyoxolane-2-carboxamide
- Nluc
- Nano-luciferase
- Nluc-A3R
- NanoLuc-labelled A3 adenosine receptor
- PMA
- (Phorbol 12-myristate 13-acetate)
- MD
- molecular dynamic
- MM-PBSA
- Molecular Mechanics-Poisson Boltzmann Surface Area
- SBDD
- structural-based drug design
- vdW
- van der Waals.