NanoBRET binding assay for histamine H2 receptor ligands using live recombinant HEK293T cells

Fluorescence/luminescence-based techniques play an increasingly important role in the development of test systems for the characterization of future drug candidates, especially in terms of receptor binding in the field of G protein-coupled receptors (GPCRs). In this article, we present the establishment of a homogeneous live cell-based BRET binding assay for the histamine H2 receptor with different fluorescently labeled squaramide-type compounds synthesized in the course of this study. Py-1-labeled ligand 8 (UR-KAT478) was found to be most suitable in BRET saturation binding experiments with respect to receptor affinity (pKd = 7.35) and signal intensity. Real-time kinetic experiments showed a full association of 8 within approximately 30 min and a slow dissociation of the ligand from the receptor. Investigation of reference compounds in BRET-based competition binding with 8 yielded pKi values in agreement with radioligand binding data. This study shows that the BRET binding assay is a versatile test system for the characterization of putative new ligands at the histamine H2 receptor and represents a valuable fluorescence-based alternative to canonical binding assays.

The histamine H 2 receptor (H 2 R), which is activated endogenously by the biogenic amine histamine (1, Fig. 1), is a long known member of rhodopsin-like receptors (class A), the largest and best studied group of G proteincoupled receptors (GPCRs) [1][2][3][4] . It represents an established target for the treatment of gastroesophageal reflux disease (GERD) and peptic ulcer, with H 2 R antagonists, like cimetidine, ranitidine and famotidine (2-4, Fig. 1) being some of the first blockbuster drugs on the market in the 1970s 5 . Current research on CNS-penetrating H 2 R ligands, especially agonists, are ongoing, to get a better understanding of the role of the H 2 R in the brain, as little is known about that so far 6 . Since the H 2 receptor has been described as being located in postsynaptic neurons and being involved in cognitive processes, it is discussed that stimulation of neuronal H 2 Rs could have similar positive effects on memory and learning as antagonizing the H 3 R 7-9 , which makes the H 2 R an interesting target for future drug development.
One of the first steps in developing novel ligands is the investigation of binding properties at the respective target. Until now, the characterization of potential ligands in terms of receptor binding is mostly done performing radioligand binding experiments. Despite its high sensitivity, the use of radiolabeled substances is usually connected with some drawbacks. In addition to the constantly increasing costs, the availability as well as the quality of commercial radioligands often declines. Furthermore, the management of radioactive waste is becoming increasingly regulated and expensive. To circumvent these issues, techniques using fluorescently labeled ligands like flow cytometry and the recently described BRET-based binding assay 10,11 , which has been adapted to several G-protein coupled receptors (GPCRs) [12][13][14][15][16][17][18][19][20][21][22][23][24] , have gained great importance.
For the NanoBRET binding assay the NanoLuc, a genetically engineered luciferase 25 , is fused to the N-terminus of the GPCR of interest 10 . After addition of the substrate the enzyme catalyzes an oxidation reaction, which is accompanied by the emission of blue light. Once a suitable fluorescent ligand binds to the tagged receptor, the ligand fluoresces due to bioluminescence resonance energy transfer (BRET). This transfer can only occur when the ligand is in close proximity to the bioluminescent donor, resulting in the observation of a lower nonspecific binding as mainly the receptor-bound fraction of the fluorescent ligand is detected. Additionally, the Scientific Reports | (2020) 10:13288 | https://doi.org/10.1038/s41598-020-70332-3 www.nature.com/scientificreports/ ligand binding process can be followed in real time and not only after equilibrium is reached, which gives deeper insight into the kinetic behavior of the ligand. In this study we established a BRET-binding assay for the histamine H 2 receptor. Therefore we synthesized three differently labeled fluorescent ligands (8-10, Fig. 1), structurally derived from BMY-25368 (5, Fig. 1), a potent and long-acting histamine H 2 receptor antagonist developed by Brystol-Myers in the 1980s 26 , and radioligand [ 3 H]UR-DE257 (7, Fig. 1) from our laboratory 27,28 . These fluorescent tracers were tested for their suitability in the BRET binding assay.

Results and discussion
Synthesis of the fluorescent ligands. The synthesis of precursor UR-DE36 (6, Fig. 1) was carried out as previously reported in five steps 28,29 . Subsequently, 6 was treated with the respective labeling reagent (13)(14)(15)Fig. 2) in the presence of triethylamine resulting in 8-10. Whereas 14 and 15 were commercially available, 13 was synthesized as described 30 . Except for 8, aminolyses worked with good yields. Analytical characterization ( 1 H-NMR, HPLC purity) of the fluorescent ligands 8-10 is shown in the Supplementary Figures S1-S6.
Properties of synthesized ligands. The chemical stability of the fluorescent H 2 R ligands 8 (UR-KAT478), 9 (UR-KAT515) and 10 (UR-KAT514) was investigated under assay conditions (pH = 7.4) in binding buffer (BB; composition see Supplementary Information) at room temperature (rt) (Fig. 3, A, B and C1). Compounds 8 and 9 showed no decomposition during 96 h (Fig. 3, A) and 24 h (Fig. 3, B) incubation, respectively, and exhibited excellent chemical stability. The stability test with compound 10 also showed no chemical degradation. However, after only one hour almost no signal corresponding to compound 10 was detectable (Fig. 3, C1), most probably because of adsorption of the ligand to the plastic vessel. This phenomenon could be confirmed visually by staining of the vessel wall and discoloration of the analyte solution. Addition of DMSO to the binding buffer (BB/DMSO 1:1) reduced adsorption, which is depicted in Fig. 3 ). Moreover, we investigated the ligands in a functional test system using a split-luciferasebased β-arrestin2 recruitment assay 31 . All ligands showed antagonistic behavior at the H 2 R (see Supplementary  Figs. S10, S11) and the obtained pK b values supported the findings from the binding assays described above (Table 1).    Fig. 4, binding was saturable for all compounds and equilibrium dissociation constants (K d /pK d values, respectively) in the nanomolar range could be determined (cf. Table 1). In congruence with the results from the other binding assays, highest affinity was observed for the BODIPY 630/650 (BY630/650)-labeled ligand 10 with a pK d of 8.59, followed by 8 (Py-1, pK d = 7.35) and 9 (TAMRA) with a pK d of 6.84 (Table 1). However, beside the moderate signal-to-background ratio of 10, a higher BRET ratio was found for 8 and 9, which makes them more suitable as BRET acceptors (Fig. 4) for screening purposes. Due to its higher binding affinity, when compared to 9, the Py-1-labeled fluorescent ligand 8 was used for further experiments.

Real-time kinetic experiments with 8.
For the further characterization of 8, real-time kinetic binding experiments were conducted (Fig. 5). Therefore, 50 nM of 8 were used to measure ligand association to the H 2 R (Fig. 5, left). The ligand was fully bound to the receptor after approximately 30 min. Dissociation of 8 was initiated by the addition of a 300-fold excess of famotidine (c = 15 µM) after preincubation (60 min) of the cells with fluorescent ligand (c = 50 nM; Fig. 5, right). Slow dissociation kinetics with a dissociation half-life of 300 min were observed and only a small amount of 8 was displaced within 240 min (35-40%). A similar behavior was also reported for the structurally related radioligand 7, leading to the assumption that the pharmacological scaffold is responsible for this type of binding 28 . All kinetic parameters describing the binding of 8 are summarized in Table 2.

Investigation of reported H 2 R ligands in BRET-based competition binding.
To show the versatility of the presented assay principle, we additionally performed BRET-based competition binding experiments with different reported H 2 receptor agonists and antagonists, using one fixed concentration of 8 (c = 50 nM) and various concentrations of the respective ligands (Fig. 6). Despite the slow dissociation kinetics of 8, all ligands were able to totally displace the fluorescent tracer after 60 min. It is noteworthy that the displacement curve of histamine Table 1. Pharmacological characterization of the fluorescent ligands 8-10 at the hH 2 R in binding and functional assays. Data represent mean values ± SEM from N independent experiments, as stated in Table 1, each performed in triplicate. a NanoBRET binding experiments performed at live HEK293T cells stably expressing the NLuc-hH 2 R. b Radioligand competition binding experiments with [ 3 H]UR-DE257 (7) (hH 2 R, K d = 11.2 nM, c = 20 nM, for representative radioligand saturation binding cf. Supplementary Fig. S7) on membrane preparations of Sf9 insect cells expressing the hH 2 R-Gsα s fusion protein as described in the Supplementary Information. c Flow cytometric measurements performed at HEK293T-hH 2 R-qs5-HA cells as described in the Supplementary Information. d β-arrestin2 recruitment assays performed at HEK293T-ARRB2-H 2 R cells as described in the Supplementary Information.  www.nature.com/scientificreports/ shows a markedly flatter slope (slope ± SEM = − 0.55 ± 0.03, N = 5) in comparison to the other tested competitive ligands, which could suggest the existence of a second receptor affinity state, as previously described 32 . However, this could not be clearly confirmed in a competition binding experiment using an extended set of concentrations (cf. Supplementary Fig. S12). Therefore, monophasic binding was assumed for all tested compounds. The pK i values from the BRET competition binding assay are shown in comparison to radioligand binding data are shown in Table 3. Data reported for radioligand binding at CHO hH 2 R membranes 32 are in good accordance with our NanoBRET data obtained at live recombinant HEK293T cells, while data acquired at Sf9 membranes expressing the hH 2 R-Gsα s fusion protein 28 show a larger deviation. It is conspicuous that agonists (1,(11)(12) show comparatively higher affinities at Sf9 membranes, whereas antagonists/inverse agonists (2)(3)(4) show lower affinities (cf. Table 3). A possible explanation for this observation could be the direct fusion of the receptor with the Gsα s , since the receptor is thereby permanently brought into an active receptor conformation favoring agonist binding. In contrast, antagonists and especially inverse agonists do not prefer this receptor state, which may lead to the observation of lower binding affinities. This is relevant as cimetidine (2), ranitidine (3) and famotidine (4) are also often described as inverse agonists at the hH 2 R, supporting our finding [33][34][35] . Another possibility for the evident discrepancy in the affinity for histamine (1) is the known allosteric effect of sodium on   www.nature.com/scientificreports/ the binding of agonists to several GPCRs 36 . As the buffer used in radioligand binding assays at Sf9 membranes expressing the H 2 R is devoid of sodium ions, in contrast to the BRET assay and also to the binding experiments performed at CHO membranes 32 , we changed the assay procedure by adding sodium in a physiological concentration (c = 145 mM) to the binding buffer. This change resulted in a decrease in affinity for histamine (1) with a pK i of 4.37 ± 0.02 (N = 3, cf.  (1), which is now in good agreement to the binding constant from the BRET binding assay. Taken together the presented BRET-based approach yields comparable binding data for reported histamine H 2 receptor ligands, which confirms the suitability of the test system in combination with the fluorescent ligand 8.

Conclusion
In this study we report the development of a NanoBRET binding assay for the histamine H 2 receptor including the synthesis and characterization of suitable fluorescent H 2 R ligands. As a homogeneous live cell-based assay, this assay allows for a convenient determination of affinity constants of putative H 2 receptor ligands, independent of their mode of action without any washing or separation steps. The results from our BRET binding assay were well comparable to currently used radioactivity-or fluorescence-based (e.g. flow cytometry) binding assays. Furthermore, real-time kinetic measurements can be performed enabling a better resolved monitoring of ligandreceptor interactions. Prerequisite for the establishment of such assays is the availability of suitable fluorescent ligands. Therefore, we synthesized three differently labeled compounds, all of which have proven to be generally usable in BRET saturation binding experiments. Out of those, substance 8 turned out to be the best compromise with regard to receptor affinity and signal strength and was successfully used for further investigations. Until now, BRET binding assays have only been described for the histamine H 1,3,4 receptors 22,23 , making this study close the gap of NanoBRET assays within the histamine receptor family. Thus, selectivity studies, which are essential for the development of new drug candidates, can be carried out using the same assay principle increasing the comparability of results. All in all, this study shows that the BRET binding assay is a valuable test system for the histamine H 2 receptor and provides a novel fluorescence-based alternative to other conventional binding assays.
Synthesis and analytical data. General  capacity factor k = (t R − t 0 )/t 0 . Absorbance was detected at 220 nm. Purity of the compounds was calculated as the percentage peak area of the analyzed compound by UV detection at 220 nm. The purities of the fluorescent ligands used for pharmacological investigation were ≥ 95%.

Scientific Reports
| (2020) 10:13288 | https://doi.org/10.1038/s41598-020-70332-3 www.nature.com/scientificreports/ Cell culture and transfection. All cells were routinely cultivated in DMEM + 10% FCS in a water-saturated atmosphere containing 5% CO 2 at 37 °C and regularly monitored for mycoplasm infection using the Venor GeM Mycoplasma Detection Kit (Minerva Biolabs, Berlin, Germany). In order to generate stable transfectants, HEK293T wild-type cells were seeded at a density of 3 × 10 5 cells/mL in a 6-well plate (Sarstedt, Nümbrecht, Germany) one day prior to transfection with 2 µg of cDNA using XtremeGene HP transfection reagent (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's protocol. After two days of incubation (water-saturated atmosphere, 5% CO 2 , 37 °C), transfected cells were trypsinized, transferred to a 15 cm cell culture dish (Sarstedt, Nümbrecht, Germany) in DMEM and geneticin was added at a final concentration of 1 mg/ mL to select for stable transfectants. After stable growth was observed, concentration of geneticin was reduced to 600 µg/mL.
NanoBRET binding assay. The cell line stably expressing the NLuc-hH 2 R construct was detached from the cell culture dishes after reaching ~ 80% confluency by treatment with trypsin/EDTA (0.05%/0.2%, for 2 min at 37 °C) and was centrifuged (600×g, 5 min). The cell pellet was then resuspended in Leibovitz' L-15 medium (L-15), supplemented with 5% FCS + 10 mM HEPES, and 1.0 × 10 5 cells/well were seeded in 70 µL (saturation and competition binding) or 80 µL (kinetic experiments) of assay medium into white 96-well cell-Grade™ plates (Brand GmbH & Co. KG, Wertheim, Germany). The cells were then incubated at 37 °C in a humidified atmosphere (no additional CO 2 ) overnight. For saturation binding experiments, serial dilutions (tenfold concentrated) of the fluorescent ligands (8)(9)(10) and famotidine (4, 300-fold excess over the respective concentration of fluorescent ligand, non-specific binding) were prepared in dilution buffer (L-15 + 2% BSA + 10 mM HEPES). 10 µL of the fluorescent ligand dilution and 10 µL of L-15 (total binding) or the dilution of 4 (non-specific binding) were added to the cells. After 60 min incubation time at 27 °C, 10 µL of the substrate furimazine, which was diluted according to manufacturer's protocol beforehand, were added. After 5 min of equilibration time at 27 °C, the measurement was started. Competition binding experiments were performed as described above using one fixed concentration of fluorescent ligand 8 (c = 50 nM) and varying concentrations of the competitors 1-4, 11, 12, that were added at the same time. Kinetic measurements were performed as follows: 10 µL of L-15 (for total binding) or 4 (300-fold excess, c = 15 µM, non-specific binding) were added to the cells. After addition of the diluted substrate, the plate was placed inside the reader for 5 min to equilibrate. To start association 50 µL of a threefold concentrated solution of the fluorescent ligand 8 (c final = 50 nM) were added to the adherent cells and the plate was measured for 60 min. Dissociation experiments were conducted in wells, which have been preincubated with 8 as described above for association experiments. To initiate dissociation, 50 µL of a fourfold concentrated solution of 4 (300-fold excess, c final = 15 µM) were added to the cells and the measurement was performed for 4 h. All measurements were performed on a TECAN InfiniteLumi plate reader (TECAN, Grödig, Austria) at 27 °C using the Blue2 NB (460 nm ± 35 nm, bandpass) and the Red NB (> 610 nm, longpass) filter combination with an integration time of 100 ms. For the kinetic experiments, integration time was increased to 500 ms for both channels to reduce noise. BRET ratios were calculated by dividing the acceptor emission (red NB) by the donor luminescence (Blue2 NB). For all BRET binding experiments, specific binding was calculated by subtracting non-specific binding from total binding yielding the "corrected BRET ratio". For saturation binding experiments, total and non-specific binding were fitted simultaneously using the "one-site total and nonspecific binding" fit. Specific binding was fitted accordingly applying the "one-site specific binding" fit. For competition binding experiments, data were normalized to buffer control (0%) and a 100%-control containing solely fluorescent ligand. The normalized competition binding curves were then fitted with a four-parameter logistic fit yielding pIC 50 -values. These were transformed into pK i -values using the Cheng-Prusoff equation 38 .
For kinetic experiments, association experiments were fitted with the "one-phase association" fit yielding k obs , whereas the dissociation experiments were fitted assuming a "one-phase decay" model resulting in k off . The association rate constant k on was calculated using the following equation: k on = (k obs − k off )/c(ligand); c(ligand) = 50 nM. Dissociation half-life t 1/2 was calculated applying the following formula: t 1/2 = ln(2)/k off .The kinetic equilibrium dissociation constant K d(kin) was calculated as follows: K d(kin) = k off /k on . The errors for k on, t 1/2 and K d(kin) were calculated according to the Gaussian law of error propagation. All experimental data were analyzed using Prism 8 software (GraphPad, San Diego, CA, USA).

Data availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.