A Radiolabeled Photoswitchable G Protein-Coupled Receptor Antagonist Enlightens Ligand Binding Kinetics Associated with Photoswitching

Abstract

Photopharmacology offers powerful opportunities to control protein signaling using photoresponsive ligands. Despite the vast potential of photoswitchable ligands for spatiotemporal target protein control, research on ligand-protein binding kinetics of these ligands remains limited. Herein, we describe the discovery of the first radiolabeled photoswitchable ligand, [³H]­VUF26063 ([³H]3f), to assess light-dependent ligand-protein binding kinetics in real time. The key compound (3f) is an arylazopyrazole-based antagonist targeting a prototypic family A G protein-coupled receptor (GPCR), the histamine H₃ receptor (H₃R), and enabled convenient radiolabeling via a growth vector on the pyrazole. Its photochemical properties, subnanomolar affinity of the trans isomer and a 50-fold decrease in affinity upon switching, allowed for reversible photochemical control of H₃R binding kinetics in real time. The kinetic binding data obtained with this radiolabeled ligand indicate that 3f isomerizes in the H₃R extended binding pocket upon illumination. Our results shed light on the binding kinetics of photoswitchable ligands and will have relevance beyond GPCRs as targets.

Introduction

Photopharmacology offers a powerful approach to regulate protein function with high spatial and temporal precision through the use of photoresponsive ligands (e.g., photocaging and photoswitching). − A variety of photoswitchable small-molecule ligands have been developed for various signaling proteins or enzymes, and are designed to reversibly isomerize between two distinct photostationary states (PSS) upon illumination, each capable of differentially modulating biology. , While equilibrium properties such as PSS composition and target affinity have been the primary focus of most studies with photoswitchable ligands, kinetic parameters and particularly ligand-protein binding kinetics have received comparatively little attention. −

In this study, we addressed this gap using the histamine H₃ receptor (H₃R), an archetypal class A G protein-coupled receptor (GPCR), as a model system. The photopharmacology of GPCRs, one of the most extensively studied protein families and the target of approximately 36% of approved therapeutics, has seen remarkable advances in recent years and therefore these targets constitute an excellent model protein family. − The H₃R is predominantly expressed in the central nervous system, where it modulates the release of histamine and other key neurotransmitters. It is implicated in various neuropathologies, including Parkinson’s and Alzheimer’s diseases, neuropsychiatric conditions such as Tourette’s syndrome and schizophrenia, and sleep disorders like narcolepsy. − Its therapeutic importance is evident from the approval of inverse agonist pitolisant for the treatment of narcolepsy. −

We have previously developed first-generation azobenzene-based H₃R antagonists (e.g., VUF14862) and agonists (e.g., VUF15000). Indeed, our 2021 in-depth analysis of the chemical diversity of photoswitchable ligands for class A GPCRs showed that azobenzenes were predominantly used as photoswitchable moieties (84%). Yet, azobenzenes exhibit well-described drawbacks, such as low aqueous solubility, dependence on ultraviolet (UV) light for trans/cis isomerization, and suboptimal PSS _trans values. Hence, improved tools for advanced studies are eagerly awaited and the photopharmacology field at large has shifted toward next-generation photoswitchable moieties to improve PSS _trans values, increase solubility and enable photoisomerization at longer wavelengths for better tissue penetration and lower tissue damage. − Among these, arylazopyrazoles have recently emerged as superior alternatives to azobenzenes, offering highly efficient and reversible switching at longer wavelengths, greater PSS percentages in both directions, and enhanced aqueous solubility owing to their heteroatom-rich structure. ,−

Using these recent advancements in the field, we developed an arylazopyrazole-based second-generation series of photoswitchable H₃R ligands with improved physicochemical, photochemical, and pharmacological properties compared to our earlier H₃R antagonist series. Key compound 3f binds H₃R with subnanomolar affinity, undergoes efficient and quantitative photoisomerization between trans and cis states and exhibits a 50-fold decrease in H₃R affinity upon switching. Importantly, the pyrazole growth vector enabled straightforward radiolabeling of 3f, yielding [³H]3f which is, to our knowledge, the first radiolabeled photoswitchable ligand. This unique and novel tool allowed us to probe the role of ligand-protein binding kinetics in light-mediated regulation of GPCR function and provided compelling evidence that photoisomerization can occur within the binding pocket of a target protein.

Results and Discussion

Design

Compound 1 was chosen as a starting point (Figure and Table ). The generally accepted pharmacophore model for nonimidazole H₃R antagonists entails a basic N,N-dialkylamino moiety separated by an aliphatic linker to an aromatic core decorated with additional substituents. , Indeed, 1 comprises a piperidine bound to an unsubstituted azobenzene with a classical aminopropyloxy linker. − To increase the affinity for H₃R, the propyloxy side chain was rigidified through a trans-cyclobutyl group (2). Rigidification is a general strategy to improve target binding affinity and has proven to be successful in the H₃R field. ,− The azobenzene was subsequently replaced by a trimethylarylazopyrazole moiety (3a). In addition to its known benefits (vide supra), the pyrazole has the additional advantage of potentially improving the required molecular recognition event between ligand and protein, as heteroatoms are often involved in productive interactions with proteins, and presenting a synthetically tractable growth vector for appending additional protein recognition elements. We pursued growing from the nitrogen atom of the pyrazole with specific groups (3b–f), aiming to further improve solubility by increasing the dipole moment, probe additional molecular interactions within the H₃R pocket, , and potentially increase the affinity shift by inducing clashes between one of the photochemical isomers and the protein. To achieve this, relatively large and rigid substituents, i.e., six-membered rings, were chosen. A substituent without a hydrogen bond donor or acceptor in the six-membered ring (cyclohexyl analogue 3b) and substituents with hydrogen bond donors and/or acceptors such as a tetrahydropyran, piperidine and N-methylpiperidine moiety (3c,e,f respectively) were introduced. The secondary amine in 3e also provided a vector that could be used for radiolabeling. The designed compounds were prepared by 3 to 4 steps via routine synthetic sequences (Scheme S1 and Figure S1) using amongst others, a Boc-protected intermediate (3d) as key precursor.

1.

Design strategy toward high-affinity arylazopyrazole-based H₃R antagonists. Improvements listed are relative to azobenzene ligand 1.

1. Photochemical Parameters, H₃R Affinity (pK _i) Values and Corresponding Affinity Shifts between PSS _cis and trans States.

^aDetermined at 25 μM in HBSS buffer containing 1% DMSO.

^bPhotostationary state area percentages after illumination of trans isomer with 365 ± 10 nm at 1 mM in DMSO for 10 min as determined by LC-MS analysis at the isosbestic point, or illumination of PSS _cis states with 506 ± 20 nm (for 3) or 434 ± 20 nm (for 1 and 2) at 1 mM in DMSO for 10 min as determined by LCMS analysis at the isosbestic point.

^cApproximate thermal relaxation half-life times of PSS _cis states in HBSS buffer + 1% DMSO, as estimated by the method of Ahmed et al. by extrapolating to 20 °C. Arrhenius plots are available in Figures S2–S7.

^dAffinity (pK _i) values as obtained from [³H]­NAMH competition experiments on cell homogenates overexpressing H₃R. Values are means of n = 3 experiments, performed in triplicate ± SEM.

^eFold shift calculated as (high-affinity state)/(low-affinity state), with the arrow indicating lower (↓) or higher (↑) affinity of the PSS _cis state in comparison to the trans state.

^fData reported by Hauwert et al.

Structure–Photochemistry and Structure–Activity Relationships

Photochemical properties and target binding affinities were monitored in parallel to guide the design and synthesis cycles (Table ). As expected, rigidification of the side chain of 1 improves target interaction, with 2 having nanomolar affinity and a 4-fold difference in affinity between trans -2 and 2 at PSS _cis (“cis-on”). Replacing the azobenzene in 2 with a trimethylarylazopyrazole (3a) results in a more than 10-fold improved K _i value for trans-3a compared to trans- 2. Since 3a at PSS _cis shows reduced affinity compared to 2 at PSS _cis , 3a has an improved photochemically induced affinity shift but, interestingly, with a trans-on profile. Next to subnanomolar H₃R affinity, 3a also shows improved (photo)­chemical properties compared to 2 and these were, as expected, also observed for its analogues 3b,c,e,f. First, as reported for many arylazopyrazole-based photoswitchable ligands, , the trans-arylazopyrazole moiety in 3a–c,e,f can be efficiently isomerized upon illumination (PSS _cis = 88–99%). Second, the estimated thermal half-life of PSS _cis states of 3a–c,e,f at 20 °C in HBSS buffer (1% DMSO) is increased from 2 days for 2 up to at least a week (Figure S2–S7, and Table ). Third, the enhanced band separation of the arylazopyrazoles , allows the use of 506 nm (green light) for 3b,c,e,f instead of 434 nm (blue light) required for 2, to more efficiently (92–95 versus 74% for 2) switch the arylazopyrazole-based ligands back from the PSS _cis state to the PSS _trans state. This observation is in line with previously reported differences in the efficiency of back-switching of azobenzene- and arylazopyrazole-based photoswitchable ligands.

While, as expected, the photochemical properties of 3a–c,e,f are not greatly affected by substitution on the pyrazole, the molecular interaction with the H₃R binding pocket is sensitive to N-substitution of the pyrazole (Figure S8 and Table ). Introducing a lipophilic cyclohexyl ring (3b) results in a 30-fold decrease in H₃R affinity compared to 3a (Figure S8A,B). Interestingly, incorporating an oxygen or nitrogen atom in the cyclohexyl ring (3c,e,f) restores high-affinity target binding in combination with an improved light-induced affinity shift of 20–50 fold (Figure S8C,D and Table ). As mentioned, (substituted) arylazopyrazole moieties potentially also improve solubility compared to azobenzene-based ligands. Therefore, nephelometry , was used to qualitatively compare the solubility of arylazopyrazole 3f to our first-generation azobenzene-based antagonist VUF14862 (Figure S9). Neither of these compounds form aggregates nor precipitate at concentrations of relevance in our pharmacology assays (vide infra). However, trans-VUF14862 shows a relatively high tendency to aggregate/precipitate at high concentrations (>10^–4.5 M), while this was not observed for 3f in either state, suggesting an improved solubility profile of 3f. Ligand 3f was selected for in-depth studies given its excellent switching properties (PSS _cis = 98%, PSS _trans = 93%), improved solubility, subnanomolar H₃R affinity (pK _i = 9.3 ± 0.1) of the trans isomer, and the 50-fold difference in affinity between trans and PSS _cis states. Its clean and robust photoswitching was further confirmed by liquid chromatography–mass spectrometry (LC–MS), NMR and UV studies (Figures S10–S12).

Binding Mode of 3f

Molecular modeling studies were performed to rationalize the observed affinities of the different isomers of 3f. Both isomers were docked in the crystal structure of the human H₃R in complex with antagonist PF03654746 (PDB: 7F61). Four independent 1 μs all-atom molecular dynamics (MD) simulations were performed for each complex, to validate predicted binding poses. The predicted pose of trans-3f (Figure A,B) is located within the orthosteric pocket with its aromatic photoswitchable moiety extending into the extended binding pocket (EBP) formed by TM2, TM7 and ECL2. This region forms a tight aromatic cluster surrounded by residues Y91^2.61, Y94^2.64, W110^3.28, Y189^ECL2, F193^ECL2, and F398^7.39. Notably, the predicted binding pose of trans -3f indicates a substantial overlap with the pose of PF03654746 in the H₃R X-ray structure (Figure S13). In contrast, the folded conformation of the cis isomer does not fit within the EBP and needs an enlarged binding pocket to be accommodated, where the arylazopyrazole moiety is located toward TM5 and 6 moving away from the EBP and its key aromatic cluster residues Y91^2.61, Y94^2.64, W110^3.28 (Figure C,D). Thus, this binding pose does not overlap well with the crystallographic binding mode of PF03654746 (Figure S13). Both trans- and cis-3f form an ionic interaction with D114^3.32, that anchors the piperidine in a similar position, deep inside the orthosteric binding pocket (Figure B,D, respectively). In the MD trajectories, the more extended, linear shape of trans-3f reaches the EBP, which results in a binding pose making increased contact with the extracellular loop region, with the N-methylpiperidine group being solvent exposed (Figure A,B). This solvent exposure can explain the structure–activity relationship (SAR) listed in Table for the arylazopyrazole-based trans isomers. While the small N-methyl group (3a) likely experiences minimal desolvation costs near the entrance of the binding pocket, the larger and apolar cyclohexyl group (3b) will experience a greater desolvation penalty. In contrast, more polar substituents (3c,e,f) are better accommodated in a solvent-exposed environment and can maintain favorable interaction with surrounding water molecules. Trans-3f also engages in aromatic interactions with Y189^ECL2, F193^ECL2, F398^7.39, an additional salt bridge with D391^7.32 and a transient H-bond interaction with Y394^7.35 (Figures B and S14A). While the folded conformation of cis-3f (Figure C) also engages in aromatic interactions with F193^ECL2, it further forms a secondary salt bridge with E395^7.36 and occasional cation-π interactions with R381^6.58 and F398^7.39 (Figures D and S14B). The conformational differences between the isomers also induce a 90-degree rotation of the cyclobutyl linker. For trans-3f, the oxygen atom in the linker is directed toward Y115^3.33, while for cis-3f, the oxygen atom is oriented toward W110^3.28, contributing to distinct spatial arrangements of the isomers within the H₃R binding site (Figure B,D). In all, the linear shape of trans-3f demonstrates an improved fit in the H₃R binding pocket compared to cis-3f, potentially explaining the observed higher binding affinity of trans-3f compared to 3f at PSS _cis .

2.

(A) Slice-through depiction of the H₃R pocket formed by trans-3f. (B) Predicted binding pose of trans-3f (green carbon atoms). (C) Slice-through depiction of the H₃R pocket formed by cis-3f. (D) Predicted binding pose of cis-3f (pink carbon atoms). Poses are based on representative MD snapshots. Ionic interactions are shown in orange dashed lines in panel B and D.

Photopharmacological Evaluation of 3f

Arylazopyrazole 3f is a H₃R ligand with subnanomolar affinity and excellent selectivity over the other three histamine receptor proteins (Table , Figures A and S15). Competition binding experiments reveal that trans- 3f does not bind to any of the other three histamine receptor subtypes at concentrations lower than 1 μM (Figure S15). Moreover, trans- 3f binds with high affinity to the mouse H₃R (pK _i = 8.8 ± 0.2) and maintains a 42-fold shift in affinity upon photoisomerization to 3f-PSS _cis (Figure S16).

3.

Photopharmacological characterization of 3f and its radiolabeled analogue [³H]3f. (A) Concentration–response curves of 3f obtained in radioligand H₃R binding assays in competition with 1.6 nM [³H]­NAMH. (B) Concentration–response curves of immethridine alone and 10 nM immethridine in competition with trans- 3f or 3f-PPS _cis in a functional H₃R-NanoBit-PKA assay after 15 min of incubation. (C) Dynamic alterations in functional H₃R-dependent nBit-cAMP sensor activity, induced by competition of 10 nM immethridine with 100 nM trans-3f and subsequent switching thereof, in comparison with immethridine activity over switching cycles. Dark indicates the preincubation time of trans-3f before immethridine stimulation and subsequent switching, magenta stripes indicate irradiation with 365 nm light, and green stripes indicate irradiation with 500 nm light. (D) Saturation binding experiments with trans-[³H]3f indicative of total, specific and nonspecific binding on cell homogenates expressing H₃R. (E) Displacement of trans-[³H]3f with H₃R-specific unlabeled ligands in radioligand binding assays. (F) Association and dissociation binding curves of trans-[³H]3f (0.5 nM). A 60 min association was performed with no competitor present. Dissociation from H₃R was initiated with 10 μM clobenpropit after 30 min of preincubation with trans-[³H]3f. (G) Light-dependent modulation of trans-[³H]3f (0.5 nM) H₃R binding and kinetics without a competitor present. (H) Dissociation of trans-[³H]3f (0.5 nM) from H₃R initiated either with 10 μM clobenpropit, or without competitor but using 365 nm light. Data represent mean ± SEM of at least n = 3 experiments, performed in triplicate. Curves are normalized to their own plateaus.

The ability of 3f to modulate H₃R signaling in living cells was evaluated. H₃R is coupled to Gα_i proteins both in native tissues and transfected cells. , The canonical inhibition of forskolin-stimulated adenylate cyclase activity following H₃R activation with the high-affinity H₃R-specific agonist immethridine (pEC₅₀ = 9.1 ± 0.1, Figure B and Table S1) was measured in transiently transfected HEK293T cells coexpressing H₃R and a NanoBiT-based cAMP sensor, measuring cAMP-dependent subunit disassembly of protein kinase A. Both trans- 3f and 3f-PSS _cis (i.e., preirradiated trans-3f) show concentration-dependent antagonism of 10 nM immethridine-induced H₃R activation (Figure B). Reminiscent of the pattern observed for 3f in radioligand binding, trans- 3f shows a significantly higher antagonistic potency (pIC₅₀ = 7.5 ± 0.1) in this functional assay, compared to 3f-PSS _cis (pIC₅₀ = 6.4 ± 0.1). The affinity determined in radioligand binding experiments shows a larger shift between trans- 3f and 3f-PSS _cis (50-fold, Table ) in comparison to the observed potency shift with the live cell cAMP sensor (12.6-fold, Table S1). This difference most likely reflects the difference in assay conditions (membranes versus live cells, different buffers). Also, bioluminescence stemming from light-based assays is known to potentially interfere with the isomeric ratio of photoswitchable ligands. ,, With nanoluciferase emitting light with a peak around 460 nm and the possibility of some 3f-PSS _cis back-switching to 3f-PSS _trans between 430 and 500 nm (Figure S10A) during the time of data collection, the observed potency shift between trans- 3f and 3f-PSS _cis might be underestimated.

Ligand 3f was investigated in an in situ dynamic photopharmacology assay using transiently transfected HEK293T cells, expressing H₃R and the cAMP sensor (Figure S17A). Preincubation of 100 nM of trans- 3f with 10 nM immethridine blocked the H₃R agonist response for approximately 80% (Figure C). Subsequent illumination with 365 nm LED light (10 min, 2 mW, Figure S17B) results in the reversal of the response up to approximately 80% of the maximal agonist response, indicative of in situ switching of trans-3f to 3f-PSS _cis and a concomitant loss of H₃R blockade. In situ back-switching with 500 nm LED light (10 min, 1.5 mW, Figure S17B) restored H₃R blockade, confirming successful 3f-PSS _cis to 3f-PSS _trans conversion. An additional 365–500 nm illumination cycle results in a similar modulation of the H₃R activity (Figures C, S17B and Table S1).

Synthesis and Application of [³H]3f

The kinetics of the molecular recognition events underlying photoswitchable ligands remain poorly understood. − Therefore, we used the desmethyl analogue 3e to prepare radiolabeled 3f (i.e., [³H]3f) as a means to directly study H₃R-ligand binding kinetics. Specifically, the availability of 3e allowed the use of a previously described CT₃ incorporation using CT₃-nosylate , (Scheme ) to prepare [³H]3f. To our knowledge, trans-[³H]3f is the first radiolabeled photoswitchable ligand ever reported. Purity and light-dependent switching of the radiolabeled compound were confirmed by liquid chromatography and the recording of [³H]-chromatograms using a β-scintillation HPLC detector. trans-[³H]3f contained 2% cis isomer, while after 10 min of irradiation with 365 nm LED light (10 V, 2 mW), 96% cis isomer was obtained (Figures S18 and S19), yielding similarly high PSS _cis values as observed for nonlabeled 3f (Table ).

1. Synthesis of [³H]3f .

^a Reagents and conditions used: (a) [³H]­methyl nosylate, MeCN, 70 °C, 15 min.

In saturation binding studies, trans-[³H]3f binds H₃R with subnanomolar affinity (pK _d = 9.2 ± 0.0, Figure D and Table ) and exhibits low nonspecific binding. The binding of trans-[³H]3f to H₃R is effectively abrogated by the reference H₃R ligands clobenpropit, thioperamide and immepip (Figure E and Table S2). Both unlabeled 3f isomers also fully abrogate trans-[³H]3f binding to H₃R, with trans-3f (pK _i = 9.0 ± 0.1) being considerably more effective than 3f-PSS _cis (pK _i = 7.7 ± 0.0, Figure E and Table S2). In kinetic experiments, the binding of 0.5 nM trans-[³H]3f to H₃R reaches equilibrium at 15 min (Figure F). Bound trans-[³H]3f dissociates from H₃R with a dissociation half-life (t _1/2) of 4.9 ± 0.4 min upon addition of an excess of clobenpropit (10 μM). This results in k _off = 0.15 ± 0.02 min^–1 and k _on = 129 ± 28 × 10⁶ M^–1 min^–1 for trans-[³H]3f. Compared to trans-[³H]3f, kinetic H₃R binding experiments with [³H]3f-PSS _cis (obtained by preilluminating trans-[³H]3f) revealed a 10-fold lower k _on value (13 ± 6 × 10⁶ M^–1 min^–1) but a similar k _off (0.11 ± 0.00 min^–1) value (Table and Figure S20A). The calculated kinetic affinities for trans-[³H]3f and [³H]3f-PSS _cis (pK _{dk in} = k _off/k _on = 8.9 ± 0.2 and 7.9 ± 0.3, respectively) correspond well to the affinities obtained in equilibrium saturation and competition binding assays (pK _d = 9.2 ± 0.0 and pK _i = 7.6 ± 0.2, respectively; Table ).

2. Kinetic Parameters of [³H]3f Binding to H₃R as Determined in Association and Dissociation Radioligand Binding Experiments, Calculated Kinetic Affinities and Affinities Obtained from Binding Experiments.

state	k on (106 M–1 min–1)	k off (min–1)	t 1/2 (min)	pK d(k in)	pK d or pK i
trans	129 ± 28	0.15 ± 0.02 (clobenpropit)	4.9 ± 0.4	8.9 ± 0.2	9.2 ± 0.0
		1.12 ± 0.28 (365 nm light)	0.8 ± 0.2
PSS cis	13 ± 6	0.11 ± 0.00 (clobenpropit)	6.2 ± 0.2	7.9 ± 0.3	7.6 ± 0.2

^aData are mean ± SEM of at least n = 3 independent experiments. Dissociation t _1/2 = ln 2/k _off.

^bpK _{d(k in)} = k _off/k _on.

^cpK _d obtained in saturation binding experiments.

^dpK _i of unlabeled PSS _cis.

Next, we probed if H₃R binding of trans-[³H]3f could be modulated by in situ photoswitching induced by illumination (Scheme S2 and Figure G). Exposure of prebound trans-[³H]3f to 365 nm light resulted in a rapid dissociation from H₃R to approximately 0% (Figure G, magenta), even though no competitor is present. This is consistent with fast quantitative photoswitching of trans-[³H]3f to [³H]3f-PSS _cis , which will not bind H₃R at this concentration due to its lower affinity. Subsequent exposure to 500 nm green light re-established H₃R binding of trans-[³H]3f over time (Figure G, green), confirming successful back-switching of the low-affinity [³H]3f-PSS _cis state to high-affinity trans-[³H]3f. Comparing the kinetics of clobenpropit-induced dissociation of trans-[³H]3f and [³H]3f-PSS _cis (Figure S20A) as well as the 365 nm light-induced dissociation of trans-[³H]3f reveals a remarkable difference in dissociation kinetics (Figures H, S20B and Table ). Photoswitching of prebound trans-[³H]3f by 365 nm illumination results in 6.2-fold faster dissociation kinetics (t _1/2 = 0.8 ± 0.2 min) as compared to the dissociation of dark trans-[³H]3f (t _1/2 = 4.9 ± 0.4 min) and [³H]3f-PSS _cis (t _1/2 = 6.2 ± 0.4 min) (Table ). Similar findings concerning light-dependent dissociation rates of photoswitchable negative allosteric modulators of the Class C GPCR mGlu₅ have been reported using a mass spectrometry binding assay. Our kinetic data exclude the dissociation process from being governed by photoswitching of unbound trans-[³H]3f in solution and subsequent readjustment of the binding equilibrium.

This photochemically induced kinetic pattern is reminiscent of the known effects of retinal confinement within a photoresponsive protein. − Retinal is a naturally occurring photoswitchable chromophore that is covalently linked to opsin proteins via a Schiff’s base to a lysine residue in TM7. Upon illumination retinal is switched from its 11-cis-retinal configuration to an all-trans-retinal form. This configurational change of the chromophore subsequently results in GPCR protein activation and triggers a cascade of biochemical events that convert light into electrical signals for vision. Retinal isomerization occurs more rapidly and with greater stereoselectivity within the binding pocket of the protein host, achieving a higher quantum yield compared to when it occurs in solution. , However, for synthetic photoswitchable compounds, including azobenzene-based ones, confinement can affect their photochemical behavior less advantegously. , As suggested for azobenzene-based ligands for an allosteric site of mGluR₅, our kinetic data provide evidence that 3f is switching when bound to H₃R. Specifically, our modeling studies (Figure ) suggest that the photoswitchable moiety of 3f sits within the EBP surrounded by flexible extracellular loops, indicating sufficient conformational flexibility to allow for this switching to occur, although the constraints of the interactions of 3f within the protein likely still reduce the efficiency of 3f to undergo a trans-cis isomerization compared to in solution. For an azobenzene-based A_2A receptor ligand, it has been proposed based on detailed spectroscopy studies that the overall outcome (efficiency and rate) of switching in the protein pocket depends on the isomerization of the azo moiety, reorientation of the phenyl group, and the potential for longer-lived excited states due to protein-coupled motion. It was also recently shown, using time-resolved crystallography, that related photoswitchable A_2A receptor ligands can be switched in the protein binding pocket and that intermediate states of ligand dissociation can be resolved. In this system, confinement of photoswitchable ligands by protein binding reduces the quantum efficiency of photoisomerization. Although we focused on binding kinetics and not on photoswitching kinetics, it is likely that the factors uncovered in the A_2A studies will also be involved in our case and will collectively manifest themselves in different binding kinetics. Indeed, the same A_2A study highlighted that photoswitching of a bound ligand can lead to conformational changes in the protein and, in line with this finding, we hypothesize that photoswitching of 3f in the protein binding pocket provides a metastable cis-3f/H₃R complex from which cis-3f is expelled faster than preformed cis-3f is expelled in the dark. As such, our data contribute to the emerging conceptual framework on the effect of protein confinement on photoswitchable ligand behavior, that is needed for the future design of optimal photoswitchable ligands.

Conclusions

We have designed and synthesized a second-generation photoswitchable antagonist series for an archetypical GPCR (H₃R), incorporating the arylazopyrazole as a photoswitchable moiety. Classical limitations of azobenzene-based photoswitchable ligands were efficiently addressed with the new series enhancing aqueous solubility, enabling the use of green light to reach PSS _trans , and boosting high PSS values both ways. Pharmacological parameters such as the overall affinity and the affinity shift were improved from template 1 to key compound VUF26063 (3f). Ligand 3f exhibits a 50-fold affinity shift between its isomers and is a high-affinity, subnanomolar trans-on antagonist for H₃R that can reversibly modulate H₃R-dependent signaling in living cells in a light-dependent manner. Capitalizing on a readily accessible synthetic growth vector in 3f, we prepared the first radiolabeled photoswitchable ligand [³H]­VUF26063 ([³H]3f). This compound was used to characterize the kinetic aspects of receptor–ligand binding events underlying the dynamic modulation of receptor signaling and these studies indicated that the trans isomer undergoes light-induced isomerization in the H₃R EBP. Our study highlights the potential of detailed kinetic characterization of the binding process for photoswitchable ligands and could aid in a better prediction of the spatiotemporal effects of protein modulation by photoswitchable small molecules.

Glossary

Abbreviations

1D NOESY

one-dimensional nuclear Overhauser enhancement spectroscopy

AC

adenylyl cyclase

BRET

bioluminescence resonance energy transfer

cAMP

cyclic adenosine mono phosphate

DMF

dimethylformamide

EBP

extended binding pocket

FBS

fetal bovine serum

H₁R

histamine H₁ receptor

H₂R

histamine H₂ receptor

H₃R

histamine H₃ receptor

H₄R

histamine H₄ receptor

HA

hemagglutinin protein tag

HBSS

Hank’s balanced saline solution

LgBit

large Bit

MD

molecular dynamics

NAMH

N^α-methylhistamine

nBit

nanoBit

nGlo

furimazine

nLuc

nanoluciferase enzyme

PBS

phosphate-buffered saline

PEI

polyethyleneimine

PSS

photostationary state

SAR

structure–activity relationship

SmBit

small Bit

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c07349.

• Synthesis of 3a-f, detailed photochemical characterization of 3f, additional modeling data on 3f, detailed experimental procedures and methods, photopharmacological characterization and chemical characterization for the compounds (PDF)

∥.

Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Jagtvej 160, 2100 Copenhagen, Denmark

§.

L.C.P.B. and I.J. contributed equally to this work.

This research was funded by the Dutch Research Council (NWO) with the grant OCENW.KLEIN.532 (“Towards the next frontiers in GPCR photopharmacology”).

The authors declare no competing financial interest.