Development of an Affinity-Based Probe to Profile Endogenous Human Adenosine A3 Receptor Expression

The adenosine A3 receptor (A3AR) is a G protein-coupled receptor (GPCR) that exerts immunomodulatory effects in pathophysiological conditions such as inflammation and cancer. Thus far, studies toward the downstream effects of A3AR activation have yielded contradictory results, thereby motivating the need for further investigations. Various chemical and biological tools have been developed for this purpose, ranging from fluorescent ligands to antibodies. Nevertheless, these probes are limited by their reversible mode of binding, relatively large size, and often low specificity. Therefore, in this work, we have developed a clickable and covalent affinity-based probe (AfBP) to target the human A3AR. Herein, we show validation of the synthesized AfBP in radioligand displacement, SDS-PAGE, and confocal microscopy experiments as well as utilization of the AfBP for the detection of endogenous A3AR expression in flow cytometry experiments. Ultimately, this AfBP will aid future studies toward the expression and function of the A3AR in pathologies.


■ INTRODUCTION
−6 The ARs are expressed on diverse cell and tissue types, in which the receptors all exert their own functions. 1−13 However, multiple contradictory observations have been reported regarding the activation of hA 3 ARs.For example, both inhibition and promotion of chemotaxis have been observed upon addition of a selective hA 3 AR agonist to neutrophils. 11,13,14Next to that, expression of the hA 3 AR is species-dependent, and large differences in hA 3 AR activity have been found between humans and rodents. 15,16Thus, many questions regarding activity and functioning of the hA 3 AR, whether on granulocytes or on other cell types and tissues, remain unanswered.
Most of the aforementioned studies have been carried out using selective ligands, e.g., agonists or antagonists to induce a cellular response as read-out.This has yielded valuable information on a biological level but ignores multiple factors that influence receptor signaling on a molecular level, such as receptor localization, protein−protein interactions (PPIs) and post-translational modifications (PTMs). 17Traditionally, these aspects would be studied using antibodies.However, antibodies for GPCRs are hindered in their selectivity due to the low expression levels of GPCRs and the high conformational variability of extracellular epitopes on GPCRs. 18This is especially true for the ARs that are lacking an extended extracellular N-terminus.
In the past decade, multiple small molecules have been developed as tool compounds to study the hA 3 AR on a molecular level.−28 Noteworthy, one of these fluorescent ligands has been used to study internalization, localization and certain PPIs of the hA 3 AR on hA 3 AR-overexpressing Chinese hamster ovary (CHO) cells as well as activated neutrophils. 12,24Yet, the current use of fluorescent ligands is limited to the type of fluorophore conjugated, a fluorescent read-out in specific assay types, and reversible binding to the receptor.Therefore, in this study, we aimed to develop a clickable affinity-based probe (AfBP) to broaden the current possibilities to measure and detect the receptor.
AfBPs are tool compounds that consist of three parts.First, an electrophilic group ('warhead') is incorporated, that facilitates covalent binding of the AfBP to the receptor (Figure 1). 32,33This allows usage of the probe in biochemical assays that rely on denaturation of proteins (e.g., SDS-PAGE and chemical proteomics).The warhead is coupled to a high to label the hA 3 AR with an AfBP.First, the AfBP is added to cells or membrane fractions to allow irreversible bond formation between receptor and probe.Click reagents are then added to install a detection moiety onto the probe-bound receptor.Lastly, cells are further processed for detection, dependent on the detection method of interest.The image of the hA 3 AR was generated using Protein Imager, 29 using the structure of the hA 3 AR (AF-P0DMS8-F1) as predicted by Alphafold. 30,31heme 1. Synthesis of hA 3 AR-Targeting Affinity-Based Probes a affinity scaffold (the second part) that induces selectivity to the protein target of interest and thirdly, a detection moiety is introduced.
Our lab has recently reported the development of electrophilic antagonist LUF7602 as an irreversible ligand of the hA 3 AR (Figure 1A). 34LUF7602 contains two out of three functionalities of an AfBP, the only part missing being the detection moiety.In the past, trifunctional AR ligands have been synthesized containing both a warhead and a detection moiety. 35,36Here, we introduced an alkyne group as the ligation handle that can be 'clicked' to a detection moiety through the copper-catalyzed alkyne−azide cycloaddition (CuAAC). 37,38This approach resulted in a 'modular' probe that can be clicked in situ to any detection moiety of interest that contains an azide group.The new probe allowed specific detection of overexpressed hA 3 AR in various assay types, such as SDS-PAGE and confocal microscopy, as well as detection of endogenous hA 3 AR in flow cytometry experiments on granulocytes.
■ RESULTS AND DISCUSSION Design and Synthesis.We decided to incorporate an alkyne group into LUF7602 analogs, to enable click-based incorporation of a detection moiety of choice onto the AfBP.Similar click strategies have already been used in the synthesis of fluorescent ligands for the hA 3 AR. 20,21,26,27,39These ligands are however all lacking the electrophilic warhead.Additionally, contrary to those studies, we mainly performed the click reaction after binding of the AfBP to the receptor, preventing a loss of affinity due to bulky substituents.−50 Also, a non-selective AfBP for the hA 3 AR has been reported, although detection of the hA 3 AR with this probe has yet to be validated. 48In previous studies on the A 1 AR, we observed that the position of the alkyne moiety on the scaffold can greatly influence the affinity of the AfBP toward the receptor, thereby affecting the functionality of the AfBP. 49To increase the chances of obtaining a successful AfBP, we therefore introduced the alkyne group on three divergent locations onto the scaffold of LUF7602 (Scheme 1).
All three synthetic routes started with compound 1, a highaffinity selective antagonist for the hA 3 AR reported over two decades ago. 51First, 1 was alkylated with bromopropane yielding tricyclic compound 2. The benzylic moiety was then removed using palladium over carbon and an excess of NH 4 HCO 2 . 34The secondary amine of 3 was alkylated with alkyne-containing fluorosulfonyl moiety 4, synthesized as recently described, 49 to yield compound 5 (LUF7930) as the first out of three AfBPs.For the second AfBP, the secondary amine of 3 was alkylated by a protected propylamine followed by deprotection of the Boc-group to yield compound 7.The methoxy group of 7 was removed using BBr 3 , yielding a zwitterionic intermediate that, after removal of remaining BBr 3 , was used immediately in an amide bond formation to synthesize fluorosulfonyl derivative 8.The alkyne moiety was then substituted onto the phenolic OH to yield compound 9 (LUF7960) as the second out of three AfBPs.For the last AfBP, the benzyl group of compound 1 was removed in the first step.However, synthesis and purification of 10 turned out to be cumbersome.Therefore, a crude mixture of 10 was used in the following alkylation step, resulting in a poor but sufficient yield of compound 11.The alkyne moiety was then substituted onto compound 11 using propargyl bromide followed by deprotection of the Boc-group to yield compound 12. Lastly, fluorosulfonyl benzoic acid was introduced using peptide coupling conditions, yielding compound 13 (LUF7934) as the final out of three AfBPs.
Affinity and Selectivity toward the hA 3 AR.The affinity of the synthesized AfBPs was determined in radioligand binding experiments.First, single concentration displacement assays were carried out on all four human adenosine receptors, using a final probe concentration of 1 μM (Table 1).Over 90% displacement was observed on the hA 3 AR, while all probes showed minimal displacement (≤25%) of the radioligand on the other ARs.This quick screen indicated a good selectivity toward the hA 3 AR over the other human ARs, a trend that was also observed in the case of the parent compound. 34Next, the 'apparent' affinities (depicted as pK i values) toward the hA 3 AR were determined using full curve displacement assays.To study the presumable covalent mode of action, the apparent affinity was determined with (pre-4 h) and without (pre-0 h) 4 h of pre-incubation of AfBP with hA 3 AR, prior to addition of radioligand.The three synthesized AfBPs showed very similar affinities at pre-0 h with apparent pK i values in the double digit nanomolar range (Table 2).In all three cases, the apparent pK i showed a strong increase upon 4 h of pre-incubation, toward values in the single-digit nanomolar range.This difference in affinity was reflected in a shift in apparent pK i .Hence, substitution of an alkyne moiety at all three of the divergent positions was well-tolerated in the case of binding affinity.To further investigate selectivity toward the hA 3 AR over the hA 1 AR, full curve displacement assays were carried out on the hA 1 AR.All AfBPs showed micromolar affinities at pre-0 h, with an increase toward submicromolar affinities at pre-4 h (Table S1).AfBP 5 showed the best selectivity toward the hA 3 AR over the hA 1 AR (>100-fold), while AfBPs 9 and 11 showed good selectivity (>17-fold and >7-fold, respectively).
To investigate the binding mode of the probes within the orthosteric binding pocket, covalent docking experiments were performed using the previously determined nucleophile Y265 as the anchor (Figure 2). 34All three compounds show a hydrogen bond interaction with the conserved N250 and π−π stacking with Phe168, two well-known interactions in ligand recognition in adenosine receptors. 52Thus, the alkyne substitution seems to be well-tolerated for all three of the probes, thereby supporting the outcome of the radioligand displacement experiments.To further confirm the covalent mode of action, wash-out experiments were performed.Compounds 5, 9, and 13 all showed persistent binding to the hA 3 AR, while full recovery of radioligand binding was observed in the case of the reversible control compound LUF7714 (Figure 3, molecular structure in Figure 1A). 34ltogether, this indicates that the three synthesized AfBPs bind covalently to the hA 3 AR.
Labeling of the hA 3 AR in SDS-PAGE experiments.Next, the potential AfBPs were investigated on their ability to label the hA 3 AR in SDS-PAGE experiments. 35,36Membrane fractions with stable expression of the hA 3 AR were incubated for 1 h with 50 nM (roughly the apparent K i ) of AfBPs 5, 9, or 13, subjected to a copper-catalyzed click ligation with sulfo-Cyanine5-azide (Cy5-N 3 ), denatured, and resolved by SDS-PAGE.Pre-incubation with the hA 3 AR-selective antagonist PSB-11 was used as a negative control, 53 and a subsequent incubation step with PNGase was introduced to remove Nglycans. 49Detection of labeled hA 3 AR turned out to be difficult if not deglycosylated: the receptor appeared as a smear at about 70 kDa (Figures 4A and 5A).Yet, this mass corresponds to the band of rat A 3 AR as has been shown in SDS-PAGE experiments on overexpressing CHO cells. 54,55N-Deglycosylation of the mixture of membrane proteins revealed a protein band at ±30 kDa in case of all three AfBPs.This protein was not labeled by any of the three AfBPs after preincubation with reversible antagonist PSB-11 and is therefore most likely the hA 3 AR.To select one of these AfBPs for further labeling studies, the intensities of the bands at ±30 kDa were compared between the probes (Figure 4B), but no significant differences were found.Correspondingly, the alkyne groups do    30,31 The extracellular side of the receptor is located at the top of the images, while the intracellular side is at the bottom.All three compounds show a hydrogen bond interaction with the conserved N250 and π−π stacking with Phe168, two well-known interactions in ligand recognition of adenosine receptors. 52The alkyne group, indicated with an arrow in each panel, fits into the binding pocket on each exit vector, and the binding orientation of the core compound, as published previously by our group, is maintained. 34ot inflict any strong unfavorable interactions upon covalent docking of the compounds in the hA 3 AR model (Figure 2).Examination of the amount of off-target labeling indicated that there are fewer other proteins labeled by 9 than by the other two probes 5 and 13 (Figure 4A).Therefore, we decided to continue our subsequent experiments with AfBP 9. Further control experiments were performed: labeling in CHO membrane fractions without expression of the hA 3 AR, no addition of probe and clicking without copper or Cy5-N 3 (Figure 4C).The band at ±30 kDa was not observed in any of the control lanes, confirming that this band is the hA 3 AR.Notably, we did not observe any hA 3 AR-specific labeling by a commercially available hA 3 AR antibody in western blot experiments (Figure S1).Presumably, the selectivity and affinity of antibodies are compromised by the relatively short N-terminus and extracellular loops of the hA 3 AR.
Labeling of hA 3 AR on Live CHO Cells.Having confirmed binding and labeling of the hA 3 AR in CHO membrane fractions, we moved toward labeling experiments on live CHO cells stably expressing the hA 3 AR.Live cells were incubated with 50 nM AfBP 9, prior to processing for either SDS-PAGE or microscopy experiments.In the case of SDS-PAGE experiments, membranes were collected, and the probebound proteins were clicked to Cy5-N 3 , denatured, and resolved by SDS-PAGE.This yielded 'cleaner' gels as compared to labeling in membrane fractions: no strong offtarget bands were observed (Figure 5).The smear of glycosylated hA 3 AR at ±70 kDa was more clearly visible as a single band (Figure 5A) and absent in the control lanes (no hA 3 AR, no AfBP or pre-incubation with PSB-11).Similar to the experiments on cell membranes, a strong reduction in size of the band was observed upon pre-incubation with PNGase (Figure 5B).Both signals were significantly reduced by preincubation with PSB-11 (Figure 5C,D).Thus, AfBP 9 also binds and labels the hA 3 AR on live cells.We speculate that the increased amount of off-target labeling in membrane fractions is due to the high enrichment of subcellular membrane proteins.Together with the electrophilic nature of the AfBP, this can result in an increased amount of protein labeling, as we have previously observed in our experiments with an electrophilic A 1 AR probe. 49n the case of the microscopy experiments, cells were fixed after probe incubation followed by a click reaction with carboxytetramethylrhodamine azide (TAMRA-N 3 ), chosen because of its uniform cellular distribution and diffusion. 56ext, multiple washing steps were carried out and cellular nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) prior to confocal imaging.A strong increase in TAMRA intensity was observed at the cell membranes upon addition of the probe to the wells (Figure 6A), visible by the increase in signal at cell−cell contacts.This signal was reduced by pre-incubation with PSB-11 and was absent for CHO cells not expressing the hA 3 AR (Figure 6B).We therefore conclude that labeling of overexpressed hA 3 AR by 9 on living CHO cells can be studied by both SDS-PAGE and confocal microscopy experiments.4 The introduction of an electrophilic warhead and clickable handle on probe 9 extends the possibilities to study the hA 3 AR in microscopy assays, e.g., by allowing workflows that are dependent on thorough washing steps and/or denaturation of proteins.
Labeling of Endogenous hA 3 AR in Flow Cytometry Experiments.Having established the potential of 9 in overexpressing cell systems, we turned to native hA 3 AR expression in flow cytometry experiments in order to cope with expected low levels of observable fluorescence caused by the notoriously low expression levels of GPCRs.Similar usage of fluorescent probes in flow cytometry experiments have thus far led to kinetic studies of ligand binding, 20,27 and detection of the hA 3 AR on the HL-60 model cell line. 21In order to establish an assay setup for primary cells we first used CHO cells as a model system.To avoid the use of excess copper on live cells, AfBP 9 was first clicked to Cy5-N 3 and desalted, before further incubation steps.Pre-clicked 9-Cy5 showed decrease in affinity toward the hA 3 AR (∼13-fold) and no binding toward the hA 1 AR in radioligand displacement assays (Table S2).9-Cy5 was then incubated for 1 h with living CHO cells with or without stable expression of the hA 3 AR.The unbound probe was removed by washing steps, and cells were analyzed by flow cytometry.The two types of CHO cells (+/− hA 3 AR) showed a difference in Cy5 mean fluorescence intensity (MFI), i.e., hA 3 AR-expressing CHO cells showed a significant increase in Cy5 MFI upon probe labeling, which was absent for CHO cells without hA 3 AR (Figure 7A,B).Next to that, pre-incubation with PSB-11 significantly reduced the observed signal in the case of the hA 3 AR-expressing CHO cells (Figure 7A,B), indicating that the observed signal is hA 3 ARspecific.Of note, no significant labeling was observed with a commercially available fluorophore-conjugated hA 3 AR antibody.
Next, we used the optimized procedure for further investigation of hA 3 AR expression on human granulocytes.Eosinophils and neutrophils were purified from human blood samples obtained from four donors, subjected to Cy5-clicked AfBP 9, and analyzed by flow cytometry.Within these experiments, purified neutrophils did not show a significant increase in Cy5 MFI upon incubation with probe (Figure 7C).The purified eosinophils however showed a significant increase in MFI that was reduced upon pre-incubation with the antagonist PSB-11 (Figure 7C,D).Thus, with the aid of AfBP 9, we were able to selectively detect hA 3 AR expression on human eosinophils, but not on human neutrophils.In the past, hA 3 AR expression has been observed on both human eosinophils and neutrophils, 7−9,12−14 though the basal hA 3 AR expression levels on human neutrophils were low in comparison to the expression on stimulated neutrophils. 12,13orrespondingly, the herein investigated neutrophils showed little to no hA 3 AR expressed on the cell surface.Future studies that make use of AfBP 9 might therefore yield new information on hA 3 AR expression levels on stimulated neutrophils, among other leukocytes.Furthermore, the role of the hA 3 AR in inflammatory conditions, such as asthma, ischemic injury, and sepsis, 58−60 might be elucidated using AfBP 9 as tool to detect receptor expression.

■ CONCLUSIONS
In this work, we have synthesized and evaluated three affinitybased probes 5, 9, and 13, which are all shown to bind covalently and with similar apparent affinities to the hA 3 AR.Although all three probes were able to label the hA 3 AR in SDS-PAGE experiments, we decided to continue with 9 due to the low number of off-target protein bands detected on gel.We showed that AfBP 9 is a versatile AfBP that can be used together with different fluorophores, examples in this study being Cy5 and TAMRA.The combination of 9 with click chemistry allowed us to label the hA 3 AR in various assay types, such as SDS-PAGE, confocal microscopy, and flow cytometry experiments, on both model cell lines and cells derived from human blood samples.In our hands, 9 showed to be successful in the selective labeling of the hA 3 AR, while commercial antibodies were not.We therefore believe that probe 9 will be of great use in the detection and characterization of the hA 3 AR in different types of granulocytes, among other cell types.
Ethics Approval.The parts of the study involving human participants were reviewed and approved by the Sanquin Institutional Ethical Committee (project number NVT0606.01).All blood samples were obtained after informed consent and according to the Declaration of Helsinki 1964.The patients and participants provided informed consent to participate in this study.
Cell Lines.Chinese hamster ovary (CHO) cells stably expressing the human adenosine A 3 receptor (CHOhA 3 AR) were kindly provided by Prof. K.N.Klotz (University of Wurzburg).CHO cells stably expressing the human adenosine A 1 receptor (CHOhA 1 AR) were kindly provided by Prof. S.J. Hill (University of Nottingham), human embryonic kidney 293 (HEK293) cells stably expressing the human adenosine A 2A receptor (HEKhA 2A AR) were kindly provided by Dr. J. Wang (Biogen), and CHO cells stably expressing the human A 2B receptor (CHO-spap-hA 2B AR) were kindly provided by S.J. Dowell (GlaxoSmithKline).
After 3 h, the filter-bound radioactivity was determined by scintillation spectrometry using a 2450 MicroBeta Microplate Counter (PerkinElmer).
Full Curve Radioligand Displacement Assay on the Adenosine A 3 Receptor.Membrane aliquots containing 15 μg of protein (CHOhA 3 AR) were resuspended in assay buffer (50 mM Tris−HCl, pH 8.0, 10 mM MgCl 2 , 1 mM EDTA, 0.01% CHAPS).The competing ligand (concentrations ranging from 0.1 to 1000 nM) was added and pre-incubated with the membrane fractions for either 0 or 4 h.The radioligand (10 nM [ 3 H]PSB-11) was then added and incubated with the samples in a total volume of 100 μL of assay buffer for 30 min at 25 °C.Nonspecific binding was determined in the presence of 100 μM NECA.Incubations were terminated by rapid vacuum filtration through prewetted 96-well GF/C filter plates using a PerkinElmer Filtermate harvester.Filters were subsequently washed 12 times with ice-cold wash buffer (50 mM Tris−HCl, pH 8.0, 10 mM MgCl 2 , 1 mM EDTA).The plates were dried at 55 °C, and MicroscintTM-20 cocktail (PerkinElmer) was added subsequently.After 3 h, the filter-bound radioactivity was determined using a Tri-Carb 2810TR Liquid Scintillation Analyzer (PerkinElmer).
Wash-out Assay.200 μL assay buffer (50 mM Tris−HCl, pH 8.0, 10 mM MgCl 2 , 1 mM EDTA and 0.01% CHAPS), 100 μL assay buffer containing the competing ligand (final concentration: 1 μM), and 100 μL of CHOhA 3 AR membrane fractions (100 μg protein) were combined and pre-incubated for 2 h at 25 °C while shaking.The '4× washed' samples were centrifuged (5 min, 13,000 rpm, 4 °C), and the supernatant was removed.The resulting pellet was dissolved in 1 mL of assay buffer and incubated for 10 min at 25 °C while shaking.The '4× washed' samples were then again centrifuged (2 min, 13,000 rpm, 4 °C), the supernatant was removed, and the resulting pellet was dissolved in 1 mL of assay buffer.The latter washing steps were repeated two times (four times washing total), and the final pellet was dissolved in 300 μL assay buffer.The radioligand (10 nM [ 3 H]PSB-11) in 100 μL of assay buffer was added to all samples, and the samples were incubated at 25 °C for 1 h.Nonspecific binding was determined in the presence of 100 μM NECA.Incubations were terminated by addition of 1 mL of ice-cold washing buffer (50 mM Tris−HCl, pH 8.0, 10 mM MgCl 2 , 1 mM EDTA) and rapid vacuum filtration through prewetted 96-well GF/B filter plates using a Brandel harvester.Filters were subsequently washed 5 times with ice-cold wash buffer.The plates were dried under vacuum, and a Micro-scintTM-20 cocktail (PerkinElmer) was added subsequently.After 3 h, the filter-bound radioactivity was determined using a 2450 MicroBeta Microplate Counter (PerkinElmer).
Data Analysis of Radioligand Displacement Assays.Data analysis was performed using GraphPad Prism version 9.0.0 (San Diego, California USA).pIC 50 values were obtained by non-linear regression curve fitting and converted to pK i values using the Cheng− Prusoff equation. 69As such, the K D values of 1.6 nM of SDS-PAGE Experiments Using Membranes.Membrane fractions were prepared as mentioned above and diluted to a concentration of 1 μg/μL.18 μL of membrane fractions from either CHOhA 3 AR or CHO-K1 cells were taken. 1 μL of competing ligand (PSB-11, unless noted otherwise) was added (final concentration: 1 μM in 0.05% DMSO in assay buffer), and the samples were shaken for 30 min at rt. 1 μL of affinity-based probe was then added (final concentration: 50 nM in 0.05% DMSO in assay buffer, unless noted otherwise), and the samples were shaken for 1 h at rt. Deglycosylation was initiated by addition of 0.5 μL of PNGaseF (5 U), 0.5 μL of MilliQ water for the control samples, and the samples were shaken for 1 h at rt.The click mix was freshly prepared by adding together 5 parts 100 mM CuSO 4 in MilliQ, 3 parts 1 M sodium ascorbate ( N a A s c ) i n M i l l i Q , 1 p a r t 1 0 0 m M T r i s ( 3hydroxypropyltriazolylmethyl)amine (THPTA) in MilliQ, and 1 part 100 μM Cy5-N 3 in DMSO.2.28 μL of click mix was added to the samples (final concentration Cy5-N 3 : 1 μM in 1% DMSO in MilliQ), and the samples were shaken for 1 h at rt. Lastly, 7.59 μL of 4× Laemmli buffer was added and the samples were shaken for at least 1 h at rt.The samples were then loaded on gel (12.5% acrylamide) and run (180 V, 100 min).In-gel fluorescence was measured on a Bio-Rad Universal Hood III using Cy3 (605/50 filter) or Cy5 (695/55 filter) settings.A Pageruler prestained protein ladder was used as the molecular weight marker.After scanning, gels were either transferred to 0.2 μM PVDF blots (Bio-Rad) using a Bio-Rad Trans-Blot Turbo system (2.5 A, 7 min) or stained with Coomassie Brilliant Blue.
SDS-PAGE Experiments Using Live CHO Cells.CHOhA 3 AR and CHO-K1 cells were cultured as mentioned above.Cells were grown to ∼90% confluency in 10 cm φ plates.10 plates were used per condition per experiment.The medium was removed, and the competing ligand (PSB-11, unless noted otherwise) in medium (final concentration: 1 μM) was added.Cells were incubated for 1 h (37 °C, 5% CO 2 ).The medium was removed, and the affinity-based probe (final concentration: 50 nM in medium) was added.The cells were incubated for 1 h (37 °C, 5% CO 2 ) and washed with PBS to get rid of all the non-bound probes.Subsequently, the membranes were prepared and collected using the procedure as described above.Membrane aliquots were diluted to a concentration of 1 μg/μL, and 20 μL was taken per sample.PNGase (0.5 μL, 5 U) or MilliQ was added, and the samples were incubated for 1 h at rt. Next, the click mix was freshly prepared by adding together 5 parts 100 mM CuSO 4 in MilliQ, 3 parts 1 M NaAsc in MilliQ, 1 part 100 mM THPTA in MilliQ, and 1 part 100 μM Cy5-N 3 in DMSO.2.28 μL of click mix was added per sample (final concentration Cy5-N 3 : 1 μM in 1% DMSO in MilliQ), and the samples were shaken for 1 h at rt. 7.59 μL of 4× Laemmli buffer was added per sample, and the samples were shaken for at least 1 h at rt.The samples were then loaded on gel (12.5% acrylamide) and run (180 V, 100 min).In-gel fluorescence was measured on a Bio-Rad Universal Hood III using Cy3 (605/50 filter) or Cy5 (695/55 filter) settings.A Pageruler prestained protein ladder was used as the molecular weight marker.After scanning, gels were either transferred to 0.2 μM PVDF blots (Bio-Rad) using a Bio-Rad Trans-Blot Turbo system (2.5 A, 7 min) or stained with Coomassie Brilliant Blue.
Western Blot Experiments.Blots were blocked for 1 h at rt in 5% BSA in TBST.Primary antibody RabbitαhA 3 AR (bs-1225R) 1:10.000 in 1% BSA in TBST was then added, and the blots were incubated overnight at 4 °C.Blots were washed (3 × TBST) and incubated for 1 h at rt with secondary antibody goatαrabbit-HRP (115-035-003) 1:2.000 in 1% BSA in TBST.The blots were washed (2 × TBST and 1 × TBS) and activated by incubating 3 min in the dark using 1 mL of both reagents of Pierce ECL Western Blotting Substrate.Blots were scanned on chemiluminescence and fluorescence using the Bio-Rad Universal Hood III.
Data Analysis of SDS-PAGE Experiments.Gel images were analyzed using Image Lab software version 6.0.1 (Bio-Rad).Quantification was done in the following manner: the area under the curve (AUC) of the bands was selected using the 'Lane Profile' tab and the adjusted volumes were taken.The adjusted volumes were then corrected for the amount of protein in each lane, using the adjusted total lane volumes of the Coomassie stained gels.The volume of the band at 55 kDa in the molecular weight marker (PageRuler Plus) was set to 100 (%), and the other bands were normalized accordingly.Further data analysis and statistics were carried out using Graphpad Prism.All given percentages are the mean values ± SEM of three individual experiments.Statistical analysis was performed using a one-or two-way ANOVA test with multiple comparisons (***p < 0.001; **p < 0.01; ns = not significant).
Confocal Microscopy.CHOhA 3 AR and CHO-K1 cells were cultured in a 96-well plate.Upon reaching a confluency of about 90%, the medium was replaced by medium containing PSB-11 (final concentration: 1 μM) or 1% DMSO (control) and the cells were preincubated for 30 min (37 °C and 5% CO 2 ).The medium was then replaced with medium containing 9 (final concentration: 50 nM) or 1% DMSO (vehicle), and the cells were incubated for 60 min (37 °C and 5% CO 2 ).The cells were washed with PBS to remove unbound probe and afterward fixed using a 4% PFA in 10% formalin solution (15 min, rt).The remaining fixative was removed by washing with PBS and subsequently with 20 mM glycine in PBS.The cells were permeabilized by incubation in 0.1% saponin in PBS (10 min, rt).Remaining saponin was removed by a PBS wash, and the cells were stored at 4 °C until further usage.At the day of imaging, PBS was removed and the cells were incubated for 1 h in freshly prepared click mix (100 μL 100 mM CuSO 4 in MilliQ, 100 μL 1 M NaAsc in MilliQ, 100 μL mM THPTA in MilliQ, 9.66 mL HEPES pH 7.4, and 40 μL 1 mM TAMRA-N 3 in DMSO, added in the respective order).The remaining click mix was removed by washing with PBS, incubating for 30 min with 1% BSA in PBS and again washing with PBS.Cells were stored in 300 nM DAPI in PBS prior to imaging.Microscopy was performed on a Nikon Eclipse Ti2 C2 + confocal microscope (Nikon, Amsterdam, The Netherlands), and this system included an automated xy-stage, an integrated Perfect Focus System (PFS), and 408 and 561 nm lasers.The system was controlled by Nikon's NIS software.All images were acquired using a 20× objective with 0.75 NA at a resolution of 1024 × 1024 pixels.The acquisition of 9 fields of view per well was done automatically using the NIS Jobs functionality.Representative images are shown in the figure and were created by using OMERO. 57ata Analysis of Confocal Microscopy Experiments.CellProfiler (version 2.2.0) was used to create a binary image of the DAPI channel and to propagate the cytoplasmic area based on the DAPI binary.An overlay of the binary cytoplasm/TAMRA channel was generated to quantify per segmented pixel the TAMRA intensity.The sum of these intensities in the cytoplasm mask is referred to as the integrated TAMRA intensity in the cytoplasm.Segmentation results were further processed using Excel while GraphPadPrism 9 was used for data visualization and statistics.
Click Reaction between 9 (LUF7960) and Cy5-N 3 .The click mix was freshly prepared by combining in a 2 mL Eppendorf tube 250 μL of 100 mM CuSO 4 in MilliQ, 150 μL of 1 M NaAsc in MilliQ, 50 μL of 100 mM THPTA in MilliQ, and 50 μL of 50 μM Cy5-N 3 in DMSO, in the same respective order.Next, 500 μL of 2 μM 9 (LUF7960) in 1% DMSO in MilliQ was added and the mixture was incubated for 1 h at rt while shaking (1000 rpm).The mixture was desalted using a Waters Sep-Pak C18 column (WAT054945).Briefly, the mixture was loaded on the column and washed three times with 1 mL of MilliQ.The product (9-Cy5) was then eluted using 1 mL of acetonitrile.The solvent was removed using an Eppendorf Concentrator Plus, and the residue was dissolved in DMSO to obtain the stock concentrations of 9-Cy5.
Flow Cytometry Experiments Using CHO Cells.CHOhA 3 AR and CHO-K1 cells were cultured as mentioned above.The cells were detached using Trypsin and centrifuged (5 min, 1500 rpm).The pellet was dissolved in medium (1 mL), the cells were counted, and the solution was diluted to 1,000,000 cells/mL.100 μL of cell solution was added to each well of a 96-well plate.The plate was centrifuged (5 min, 1500 rpm), and the medium was replaced by medium containing PSB-11 (final concentration: 1 μM) or 1% DMSO (control).Cells were resuspended and incubated for 30 min (37 °C, 5% CO 2 ).The plate was centrifuged (5 min, 1500 rpm), and the medium was replaced by medium containing pre-clicked 9-Cy5 (final concentration: 50 nM).Cells were resuspended and incubated for 1 h (37 °C, 5% CO 2 ).The plate was centrifuged (5 min, 1500 rpm), the medium was removed, and 1% BSA in PBS containing rabbitαhA 3 AR PE conjugate (1:500) was added.Cells were resuspended and incubated for 30 min at 4 °C.The cells were washed with 1% BSA in PBS to remove unbound probe and antibody and resuspended in 1% BSA in PBS.Cells were measured using a Cytoflex S (Beckman and Coulter, USA), and data was analyzed with FlowJo v10.0.7 (BD Life Sciences).The gating strategy is shown in Figure S2.
Flow Cytometry Experiments Using Human Granulocytes.100 μL of PMNs or purified eosinophils (2•10 6 cells/mL) in HEPES buffer was incubated with 50 μL of PSB-11 (final concentration: 1 μM) or 1% DMSO (control) in HEPES buffer for 30 min at rt. 50 μL of pre-clicked 9-Cy5 (final concentration: 50 nM) or 1% DMSO (vehicle) was added, and the samples were incubated for 1 h at rt.The samples were then added to a 96 well-plate and centrifuged (5 min, 1500 rpm, 4 °C).The supernatant was removed, and the pellet was suspended in a solution containing antibody (Siglec-8 PE conjugate 1:100) in 0.5% HSA in PBS.The samples were incubated for 30 min at 4 °C, centrifuged (5 min, 1500 rpm, 4 °C), and washed once with 0.5% HSA in PBS.The final pellet was suspended in 0.5% HSA in PBS.Cells were measured on a Canto II flow cytometer (BD, Franklin Lakes, NJ, USA) and analyzed with FACSDiva software.The gating strategy is shown in Figure S2.
Data Analysis of Flow Cytometry Experiments.Data analysis was performed using Graphpad Prism.MFI values shown are mean ± SEM from three individual experiments performed in duplicate (Figure 7A) or four individual experiments (Figure 7C).Statistical analysis was performed using a one-way ANOVA test with multiple comparisons (****p < 0.0001; ***p < 0.001; **p < 0.01; ns = not significant).

Figure 1 .
Figure 1.(A) Previous work: covalent hA 3 AR antagonist LUF7602 and control compound LUF7714; (B) this work: AfBP LUF7960; (C) strategyto label the hA 3 AR with an AfBP.First, the AfBP is added to cells or membrane fractions to allow irreversible bond formation between receptor and probe.Click reagents are then added to install a detection moiety onto the probe-bound receptor.Lastly, cells are further processed for detection, dependent on the detection method of interest.The image of the hA 3 AR was generated using Protein Imager,29 using the structure of the hA 3 AR (AF-P0DMS8-F1) as predicted by Alphafold.30,31

[ 3 H
]PSB-11 binding on CHO cell membranes stably expressing the hA 3 AR at 25 °C after 0.5 h of co-incubating the probe and radioligand.b Apparent affinity determined from displacement of specific [ 3 H]PSB-11 binding on CHO cell membranes stably expressing the hA 3 AR at 25 °C after 4 h of pre-incubation with the respective probe followed by an additional 0.5 h of co-incubation with the radioligand.c Fold change determined by ratio K i (0 h)/K i (4 h).d ****p < 0.0001 compared to the pK i values obtained from the displacement assay with 0 h pre-incubation of the probe, determined by a one-way ANOVA test using multiple comparisons.e Data represent the mean ± SEM of three individual experiments performed in duplicate.

Figure 2 .
Figure 2. Putative binding modes of compounds 5 (A), 9 (B), and 13 (C) in an Alphafold model of the hA 3 AR (AF-P0DMS8-F1-model_v4).30,31The extracellular side of the receptor is located at the top of the images, while the intracellular side is at the bottom.All three compounds show a hydrogen bond interaction with the conserved N250 and π−π stacking with Phe168, two well-known interactions in ligand recognition of adenosine receptors.52The alkyne group, indicated with an arrow in each panel, fits into the binding pocket on each exit vector, and the binding orientation of the core compound, as published previously by our group, is maintained.34

Figure 3 .
Figure 3. Wash-out assay reveals covalent binding of all three probes to the hA 3 AR.CHO cells membranes stably transfected with the hA 3 AR were pre-incubated with 1% DMSO (vehicle), 1 μM noncovalent control compound LUF7714, or 1 μM compounds 5, 9, or 13.The samples were washed for either 0 or 4 times, before being exposed to [ 3 H]PSB-11 in a radioligand displacement assay.Data is expressed as the percentage of the vehicle group (100%) and represents the mean ± SEM of three individual experiments performed in duplicate.****p < 0.0001 determined by a two-way ANOVA test using multiple comparisons.

Figure 4 .
Figure 4. Labeling of the hA 3 AR by the synthesized affinity-based probes.(A) Labeling of proteins by 5, 9, and 13.Membrane fractions from CHO cells stably overexpressing the hA 3 AR were pre-incubated for 30 min with the antagonist (PSB-11, 1 μM final concentration) or 1% DMSO (control), prior to incubation for 1 h with the respective probe (50 nM final concentration).The proteins were subjected to PNGase or MilliQ (control) for 1 h to remove N-glycans.Samples were then subjected to a copper-catalyzed click reaction with Cy5-N 3 (1 μM final concentration), denatured using Laemmli buffer (4×) and resolved by SDS-PAGE.Gels were imaged using in-gel fluorescence and stained with Coomassie Brilliant Blue (CBB) as protein loading control.(B) Quantification of the lower hA 3 AR band (±30 kDa).The band intensities were taken and corrected for the observed amount of protein per lane upon CBB staining.The band at 55 kDa of the PageRuler Plus ladder (not shown) was set to 100% for each gel, and band intensities were calculated relative to this band.The mean values ± SEM of three individual experiments are shown.Significance was calculated using a one-way ANOVA test using multiple comparisons (ns = not significant).(C) Control experiments with probe 9. Membrane fractions from CHO cells with or without (first lane) stable expression of the hA 3 AR were pre-incubated for 30 min with antagonist (PSB-11, 1 μM final concentration) or 1% DMSO (control), prior to incubation for 1 h with 9 (50 nM final concentration) or 1% DMSO (control).Proteins were deglycosylated with PNGase for 1 h.The click mix was then added, containing CuSO 4 or MilliQ (control) and Cy5-N 3 (1 μM final concentration) or DMSO (control).The samples were then denatured with Laemmli buffer (4×) and resolved by SDS-PAGE.Gels were imaged using in-gel fluorescence and afterward stained with CBB.The image shown is a representative of three individual experiments.

Figure 5 .
Figure 5. Labeling of the hA 3 AR on live CHO cells.CHO cells with or without (first lane) stable expression of the hA 3 AR were preincubated for 1 h with antagonist (PSB-11, 1 μM final concentration) at 37 °C, prior to incubation with 9 (50 nM final concentration) for 1 h at 37 °C.After the incubation, the unbound probe was washed away with PBS.Membranes were prepared, brought to a concentration of 1 μg/μL, and subjected to the copper-catalyzed click reaction with Cy5-N 3 (1 μM final concentration).Samples were then denatured with Laemmli buffer (4×), resolved by SDS-PAGE, and imaged using ingel fluorescence.Gels were stained by Coomassie Brilliant Blue (CBB) as loading control.(A) Labeling of glycosylated hA 3 AR.(B) Labeling of deglycosylated hA 3 AR.PNGase was added prior to the addition of click reagents.(C, D) Quantification of the observed signals with and without addition of antagonist (PSB-11).The band intensities were calculated using ImageLab and corrected for the amount of protein measured after CBB staining.The band at 55 kDa of the PageRuler Plus ladder (not shown) was set to 100% for each gel and band intensities were calculated relative to this band.The mean values ± SEM of three individual experiments are shown.Significance was calculated by a two-way ANOVA test using multiple comparisons (***p < 0.001; **p < 0.01).

Figure 6 .
Figure 6.Labeling of the hA 3 AR observed by confocal microscopy.CHO cells with (CHO-hA 3 AR) or without (CHO-K1) stable expression of the hA 3 AR were pre-incubated for 30 min with PSB-11 (1 μM final concentration) or 1% DMSO (control) and incubated for 60 min with 9 or 1% DMSO (vehicle control).Cells were fixed, permeabilized, and subjected to a copper-catalyzed click reaction with TAMRA-N 3 (1 μM final concentration).The cells were then washed and kept in PBS containing 300 nM DAPI during confocal imaging.(A) Shown are DAPI staining (blue, first row), TAMRA staining (yellow, second row), and an overlay of both stains (third row).Images were acquired automatically at multiple positions in the well of interest and are representatives from two biological experiments.Scale bar = 50 μM.Figure was created using OMERO. 57(B) Comparison of the integrated fluorescence intensity between treatment conditions.Data was obtained from 2 × 9 fields of view, from the same experiment performed in duplicate.Each data point represents the integrated fluorescent intensity of the TAMRA signal per individual cell.Shown in the bar graphs is the average integrated fluorescence intensity of all individual cells ± SEM.Significance was calculated using a one-way ANOVA test using multiple comparisons.A significant increase in intensity is observed for the cells containing the hA 3 AR and treated with 9, versus the other conditions.

Figure 7 .
Figure 7. Labeling of the hA 3 AR in flow cytometry experiments.Samples were pre-incubated for 30 min with the antagonist PSB-11 and incubated for 60 min with 9 pre-clicked to a Cy5 fluorophore (9-Cy5).Samples were then washed and analyzed on Cy5 fluorescence by flow cytometry.(A) Cy5 mean fluorescence intensity (MFI) in CHO cells with (CHO-hA 3 AR) and without (CHO-K1) stable expression of the hA 3 AR.Values represent the mean ± SEM of three individual experiments performed in duplicate.Significance was calculated by a one-way ANOVA test using multiple comparisons (***p < 0.001; **p < 0.01; ns = not significant).(B) Representative graph showing the observed shift in MFI related to hA 3 AR labeling in CHO-hA 3 AR cells.(C) MFI of 9-Cy5 in neutrophils and eosinophils purified from human blood samples.Values represent the mean ± SEM (n = 4) of four donors from two individual experiments.Significance was calculated by a one-way ANOVA test using multiple comparisons (****p < 0.0001; ***p < 0.001; ns = not significant).(D) Representative graph showing the observed shift in MFI related to hA 3 AR labeling in human eosinophils.
[ 3 H]DPCPX at CHOhA 1 AR membranes and 1.7 nM of [ 3 H]PSB603 at CHOspap-hA 2B AR membranes were taken from previous experiments, 70,71 while the K D values of 1.0 nM of [ 3 H]ZM241385 at HEK293hA 2A AR membranes and 17.3 nM of [ 3 H]PSB-11 at CHOhA 3 AR membranes were taken from in-house determinations.Single point displacement values shown are mean percentages of two individual experiments performed in duplicate.All pK i values shown are mean values ± SEM of three individual experiments performed in duplicate.Statistical analysis was performed using a one-way ANOVA test with multiple comparisons (****p < 0.0001).

Table 1 .
Radioligand Displacement (%) of the Synthesized hA 3 AR Probes on the Four Adenosine Receptors e