Development of a HiBiT Peptide-based NanoBRET Ligand Binding Assay for Galanin Receptor 1 in Live Cells

Abstract

Galanin is a neuroendocrine peptide regulating a wide range of physiological functions, including feeding and energy homeostasis, mood and anxiety, and modulation of pain. The function of galanin peptide is mediated through its three galanin receptors, GALR1, GALR2 and GALR3, which belong to the G protein-coupled receptor family. To measure the interaction of ligands with galanin receptor 1 (GALR1) in living cells, we developed a novel HiBiT peptide-based NanoBRET^™ ligand binding assay. We generated six bioluminescence resonance energy transfer (BRET) tracers comprised of modified and truncated galanin peptide derivatives tagged with a 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY) acceptor fluorophore. The fluorophore-tagged peptide tracers were evaluated in cells expressing GALR1 tagged with HiBiT, an 11-amino acid subunit of NanoLuc^®, HiBiT, which upon high affinity complementation with the cell-impermeable subunit LgBiT reconstituted a functional NanoLuc^® Luciferase. Addition of furimazine substrate induced BRET to the BODIPY fluorophore acceptor component of the galanin-derived peptide tracers and produced a fluorescent signal output. Using this BRET assay, we characterized the binding affinity and binding kinetics of tracers with GALR1 in both equilibrium and real time. To validate our assay, we evaluated the binding affinity and function of a panel of unmodified galanin-derived peptide ligands through competitive displacement of bound fluorescent galanin tracers. Our data showed that the binding affinity of these galanin peptide ligands correlated well with their rank order in β-arrestin recruitment and internalization functional assays. This study demonstrates that the HiBiT peptide-based NanoBRET^™ ligand binding assay is a valuable system for studying ligand engagement of GALR1 in living cells, offering an alternative to neuropeptide radioligand binding assays.

Graphical Abstract

Introduction

Galanin is a neuroendocrine peptide, first discovered from the porcine intestinal tract by Mutt et al. in the early 1980s.¹ Mature galanin (30 amino acids in human, 29 amino acids in rodents) is derived from a precursor protein through sequential cleavage of an N-terminal signal peptide and two pairs of basic amino acids². Galanin is widely distributed throughout the body, including the central and peripheral nervous systems as well as the endocrine system. It regulates a wide range of physiological functions such as feeding and energy homeostasis, mood and anxiety, and modulation of pain.² Galanin peptide signaling is mediated through its three galanin receptors-GALR1, GALR2 and GALR3-which belong to the G protein-coupled receptor (GPCR) family.³ Truncated, chimeric peptide ligands, as well as peptide-like ligands have been developed to study the interaction of peptide ligands and the three subtype galanin receptors utilizing radioligand competitive binding assays.^{3, 4}

There are inherent drawbacks to the radioligand binding assay despite its high sensitivity. First, the assay requires a laboratory that can handle radioactive substances and manage radioactive waste, and the high cost and availability of radioligands can be limiting. Second, these assays utilize membrane preparations, and thus they study ligand interactions with GPCRs in cell extracts rather than intact cells, which may exclude critical interactions within the live cell. Third, radioligand competitive binding assays require a washing step to remove nonspecific binding which is less amenable to kinetic measurements, especially for those ligands with fast k_on or k_off. Fourth, while some examples of low-throughput screening with radioligand-based assays exist, high-throughput screening with 384-well microplate and greater are not compatible with radioactive assays. As an alternative pharmacology approach, novel NanoBRET-based target engagement/displacement assays have been developed in recent years to investigate the interaction of GPCRs and their ligands in living cells.^5–8 Promega has pioneered the use of BRET tracers derived from promiscuous GPCR ligands to study target engagement in multiple GPCRs.^{7, 9, 10} The NanoBRET assay measures the proximity or binding of a ligand to a target protein. Rather than being radiolabeled, BRET tracers typically contain a BRET energy acceptor in the form of a chemical fluorophore that is tethered via a flexible linker to the ligand that binds to the GPCRs. The BRET energy donor, NanoLuc^™ Luciferase (NLuc), emits “donor” bioluminescence upon addition of its optimized substrate furimazine (Figure 1). Instead of the full NLuc protein, the exposed extracellular portion of the GPCR is often tagged with the smaller 11-amino acid HiBiT, which generates bright bioluminescence upon high affinity complementation with LgBiT, an 18 kDa cell impermeable subunit derived from NLuc. Small peptides like HiBiT often exert minimal influence on the native structure and function of the GPCRs, and the cell impermeability of LgBiT restricts complementation at the cell surface. Upon addition of the substrate, the functional HiBiT/LgBiT complex produces “donor” bioluminescence (Figure 1). Given the high distance constraints of resonance energy transfer (<10 nm), BRET occurs only when the fluorescent tracer binds to the HiBiT-tagged receptor and is in close proximity to the donor (<10 nm).¹¹ The assay therefore only provides signal from the engineered target of interest (excluding non-specific binding), and the high specificity of BRET allows quantification of tracer engagement through ratiometric measurements (dividing the acceptor emission intensity by the donor emission intensity).

Figure 1.

Schematic of NanoBRET-based target engagement assay. In this assay, HiBiT tag (11 amino acids) is genetically fused to the N-terminus of GALR1. HiBiT forms the HiBiT/LgBiT luciferase complex through high affinity complementation (K_D = 0.7 nM) with an 18 kDa subunit derived from NanoLuc known as LgBiT and produces bright luminescence upon substrate (furimazine) addition. When galanin-BODIPY tracer binds to the receptor, bioluminescence resonance energy transfer (BRET) occurs between HiBiT/LgBiT complex (donor) and galanin-BODIPY tracer (acceptor). The BRET signal (measured by fluorescence output) is used to quantify dynamic interactions between the unmodified ligands and GALR1 through competitive binding with galanin-BODIPY tracer.

Here we report the development of a NanoBRET ligand binding assay for GAL1R, including the optimization of new probes. Since there are no known small molecule modulators of GALR1, we chose to derivatize known peptide modulators of GALR1 for tracer development, namely, 30-mer human galanin peptide, galanin 1-15 (Gal1-15), and the 20-amino acid ligand M40.¹² Each peptide was modified with a cell-permeable, NanoBRET^™-compatible fluorophore i.e., NanoBRET^® 590, a BODIPY-based dye. The NanoBRET^™ 590 dye was chosen due to its optimal spectral overlap with NLuc for BRET. Given tight proximity constraints for effective BRET, NanoBRET^™ 590 dyes with two different linker lengths were chosen for attachment to each peptide.¹³ Altogether, six fluorescently tagged peptide tracers were synthesized and evaluated for their binding properties to HiBiT-tagged GALR1 expressed on the surface of HEK293 cells. The binding kinetics of the tracers were characterized in living cells, under both equilibrium conditions and in real time. To further validate the binding properties of tracers, we performed the radioligand binding assay using [³H]Galanin as the radioligand. In addition, to demonstrate the utility of this assay, the binding affinity of a panel of unmodified galanin-based peptide ligands was evaluated by displacing the tracers from HiBiT-GALR1 expressing cells.

Material and Methods

Peptide synthesis and fluorophore conjugation:

Peptides were synthesized using 9-fluorenyl methoxycarbonyl (Fmoc) solid-phase peptide chemistry with either high or low-loading Rink amide 4-methylbenzhydrylamine (MBHA) resin (HL, 0.64 mmol/g Lot no. 26874 or LL, 0.32 mmol/g Lot no. 27239) from Gyros Protein Technologies Inc. NanoBRET^™ 590 SE Dye (CAS: 201998-61-0) and NanoBRET^™ 590-C4-SE Dye (CAS: 2676206-06-5) were used as supplied.¹³

Method A: Non-tracer Peptide Synthesis and Cleavage

Synthesis was performed on a CEM Liberty Blue^™ automated microwave peptide synthesizer. Coupling reactions on the Liberty Blue^™ were performed under ethyl(hydroxyimino)cyanoacetate (Oxyma)/ N, N′-diisopropylcarbodiimide (DIC) conditions on a 0.1 mmol scale.

Peptides attached to the resin, with free N-terminal amine (Fmoc-cleaved), were drained from their suspension solution on a CEM Razor^® instrument and washed with DMF followed by DCM. The peptides on resin were then suspended in a deprotection cocktail (5 mL, 0.02 M) containing 88% trifluoroacetic acid (TFA), 5% H₂O, 5% Phenol, and 2% triisopropyl silane (TIPS). The resin suspension was allowed to sit for 20 min at 42 °C before collecting the supernatant. The resin was washed with DCM (2 mL x 3) with supernatant and washings combined to an approximate total volume of 11 mL. To this, cold diethyl ether (35 mL) was added resulting in the precipitation of the resin cleaved peptide, which was then centrifuged at 10,000 rpm, 4 °C for 15 min. The supernatant liquid was decanted, the solid was dissolved in DMSO (2 mL) and purified by reverse phase chromatography (RediSep^® Prep C18 Column, 20 X 250 mm, 100 Å / 5μm, gradient of 10% to 100% ACN (0.1% TFA) in H₂O (0.1% TFA). Fractions were analyzed by LC/MS and those containing pure desired product were combined and subjected to lyophilization to produce each peptide as a white fluffy solid.

Method B: Tracer Peptide Synthesis and Cleavage

Synthesis was performed on a CEM Liberty Blue^™ automated microwave peptide synthesizer with a modified method toward the generation of tracer peptides. Coupling reactions on the Liberty Blue^™ were performed under ethyl(hydroxyimino)cyanoacetate (Oxyma)/ N, N′-diisopropylcarbodiimide (DIC) conditions on a 0.1 mmol scale. However, the final Fmoc deprotection step was omitted from this sequence. Fmoc-protected peptides on resin were processed according to method A to produce pure resin-cleaved, Fmoc-protected peptides as white fluffy solids.

NanoBRET^™ 590 Peptide Conjugation

General Procedure A

To Fmoc-protected tracer peptide precursor (1 equiv.) in an oven-dried, N₂ flushed 20 mL scintillation vial was added DMF (500 μL - 1 mL) and the solution was allowed to stir at room temperature for 5 min. To this, DIPEA (5 equiv.) was added, and the mixture was stirred for 5 min. Separately, an oven-dried 2-dram glass vial containing the appropriate NanoBRET^™ 590-SE dye (1 – 1.05 equiv.) was flushed with N₂ for 5 min. To this vial, DMSO (500 – 700 μL) was added and the mixture was sonicated for 10 sec. The deep purple mixture was added to the reaction mixture dropwise while stirring followed by a DMSO rinse of the 2-dram glass vial/needle (320 – 625 μL). The reaction mixture was stirred at room temperature in the absence of light and reaction progress was monitored by LC/MS (4% to 100% ACN in H₂O (0.0.5% TFA). If LC/MS analysis showed the reaction progress had stalled after 1 h, additional DIPEA (2.5 - 5 equiv.) was added.

Once the coupling reaction was judged to have gone to completion, piperidine (230 – 750 μL) was added to bring the solvent to piperidine ratio to 4:1. LC/MS analysis taken post piperidine addition showed instant cleavage of the Fmoc protecting group. The crude reaction mixture was then injected directly into a RediSep Gold^® C18 30 g column and purified with a gradient of 10% to 100% ACN (0.1% NH₄OH) in H₂O (0.1% NH₄OH). Eluted fractions containing pure product as analyzed by LC/MS were combined and subjected to lyophilization to afford the desired peptide tracer as a purple fluffy solid.

Tracers 1-6 were prepared according to General Procedure A. LCMS purity and HRMS identification data are included in the Supplementary Information.

Constructs and transfection

pATG4092-HiBiT-GALR1, pATG4092-HiBiT-GalR2 and pATG4092-HiBiT-GalR3 constructs were provided by Promega. The constructs contained N-terminal IL-6 secretion tag (MNSFSTSAFGPVAFSLGLLLVLPAAFPAP) followed by a VS linker, a HiBiT tag (VSGWRLFKKIS), a 2× GSSG linker and human GALR1, human GALR2 and human GALR3, respectively.

HEK293 cells were obtained from ATCC and cultured at 37 °C and 5% CO₂ in Dulbecco’s modified Eagle’s medium (ThermoFisher Scientific) containing 10% fetal bovine serum (ThermoFisher Scientific) and 100 units/ml of penicillin–streptomycin (ThermoFisher Scientific).

The constructs were transfected into HEK293 using FuGENE^® HD Transfection Reagent (Promega) at a 1:3 DNA/transfection reagent ratio. Briefly, DNA encoding HiBiT-GALR1 fusion protein (0.1 μg) and transfection carrier DNA (9.9 μg) was diluted in 0.5 mL Opti-MEM without serum (ThermoFisher Scientific). FuGENE^® HD transfection reagent (30 μL) was diluted in 0.5 mL Opti-MEM without serum. DNA and transfection reagent were mixed and incubated at room temperature for 15 min. HEK293 cells were harvested and suspended in Opti-MEM without phenol red supplemented with 2% FBS and 100 units/mL penicillin–streptomycin at a final concentration of 2.2 x 10⁵ cells/mL. 1 mL DNA–transfection reagent complexes were combined with 9 mL HEK293 cells and dispensed in white 384-well plate (36 μL/well) or white 1536-well plate (5 μL/well) by using Multidrop Combi reagent dispensers (ThermoFisher Scientific). The cells were grown at 37 °C and 5% CO₂ for 22-24 hours.

PathHunter GALR1 β-arrestin recruitment assay

PathHunter CHO-K1 GALR1 β-arrestin cells were purchased from Eurofin/DiscoverX (Catalog# 93-0228C2). GALR1 β-arrestin cells were maintained in the growth medium consisting of 1x F-12 Nutrient Mixture (HAM) (Thermo, # 11765054), 10% FBS (Thermo, # 10082), 1xPen/Strep, 800μg/mL Geneticin (Thermo, # 10131035), and 300μg/mL Hygromycin B (Thermo, # 10687010). For performing the assay, cells were harvested with StemPro Accutase Cell Dissociation Reagent (Thermo, # A1110501), and seeded in 1536-well white flat bottom plates (Greiner, #789173-F) at the density of 1,000 cells/ 3μL/ well in AssayComplete Cell Plating Reagent 2 (Eurofin/DiscoverX, # 93-0563R2B). After overnight incubation at 37°C, 5% CO₂, 20 nL/well of peptides were added to each well with Echo 650 Acoustic Liquid Handlers (Beckman Coulter, Brea, CA). After incubation at 37°C, 5% CO₂ for 90 min, 1.5μL/ well of PathHunter detection reagent (DiscoverX# 93-0001L) was added to microplates with BioRAPTR FRD dispenser (Beckman Coulter, Brea, CA). The detection reagent was prepared by mixing together Galacton Star Substrate, Emerald II solution and PathHunter buffer (supplied by the assay kit) at 1:5:19 proportion accordingly prior to dispensing. The microplates were incubated at room temperature for 1 h and then the luminescent signal was detected on ViewLux plate reader (PerkinElmer, Waltham, MA).

BRET imaging

HeLa cells (ATCC) were plated in tissue culture treated 8-well imaging chambers (ibidi) at density of 2× 10⁴ cells / well in 200 μL of Opti-MEM without phenol red supplemented with 2% FBS and 100 units/ml penicillin–streptomycin and incubated 18–24 hours at 37 °C /5% CO₂. Next day, cells were transfected with the HiBiT-GALR1 construct at a 1:10 dilution with a promoterless carrier DNA using ViaFect transfection reagent (Promega) at a 1:3 DNA/transfection reagent ratio and grown for 18–24 hours at 37 °C /5% CO₂. For BRET imaging, cell with treated for 30 min at 37 °C /5% CO₂ with Tracer 1 at a final concentration of 1 μM in the presence or absence of 30 μM Galanin. For detection cells were treated for 10 min with 200 μL of detection reagent comprising of 100-fold dilution of LgBiT and 50-fold dilution of Nano-Glo HiBiT extracellular substrate in Opti-MEM without phenol red and then imaged. Images were taken on a GloMax^® Galaxy bioluminescence imaging system (Promega) equipped with a temperature-controlled stage. Each image is an average projection of 5 individual frames. Donor channel: 460/50 nm bandpass, Acceptor channel: 575 nm longpass. Acquisition time: 120 second donor, 120 second acceptor.

BRET assay

For the saturation binding assay, transfected cells were seeded in white 1536-well plate (5 μL/well) and incubated at 37 °C and 5% CO₂ for 22-24 hours. Tracers were serially diluted (top final concentration 3 μM, 3 times dilution) in DMSO solution and transferred by Echo 650 acoustic liquid handler (Beckman Coulter). To determine nonspecific binding of tracers, excessive unmodified galanin peptide (33 μM final concentration) was added to the control wells. The plates were mixed briefly and incubated at room temperature for 90 min. To measure BRET signal, 5 μL 2xHiBiT detection regent (consisting of 100-fold dilution of LgBiT and 50-fold dilution of Nano-Glo HiBiT extracellular substrate in Opti-MEM without phenol red) was added into each well, and the plate was shaken at room temperature for 15 min to allow HiBiT/LgBiT complementation. BRET signal was measured on a PHERAstar FSX microplate reader equipped with 450/80BP and 610/LP filters (BMG LabTech). To generate BRET ratio values (milli BRET Units, mBU), the acceptor emission value (610 nM) was divided by the donor emission value (450 nM) and multiplied by 1,000. Specific BRET was calculated by subtracting BRET values in the presence of excess unmodified galanin peptides from BRET values in the absence of galanin peptides.

For competitive displacement assay, transfected HEK293 cells were seeded in white 1536-well plate (5 μL/well) and incubated at 37 °C and 5% CO₂ for 22-24 hours. The unmodified peptide ligands were serially diluted in the DMSO solution and transferred to the cells along with EC₈₀ concentration of tracers (calculated from saturation binding data) by Echo 650 acoustic liquid handler. The plates were mixed briefly and incubated at room temperature for 90 min. 5 μL 2xHiBiT detection regent was added into cells and BRET signal was measured on a PHERAstar FSX microplate reader. K_i values were calculated according to the Cheng–Prusoff equation.¹⁴

Radioligand binding assay

HEK293 cells stably expressing human GALR1 were grown in Dulbecco’s modified Eagle medium (DMEM), supplemented with 10% fetal bovine serum, 1% of antibiotic/antimycotic, 2 mM of L-glutamine and Hygromycin B (50 mg/mL) and kept in an incubator at 37 °C and 5% CO₂. Upon reaching 80–90% confluence, cells were harvested using premixed Earle’s balanced salt solution with 5 mM ethylenediaminetetraacetic acid (EDTA) (Life Technologies) and centrifuged at 3000 rpm for 10 min at 21 °C. The supernatant was removed, and the pellet was resuspended in 10 mL of hypotonic lysis buffer (5 mM MgCl₂, 5 mM Tris, pH 7.4 at 4 °C) and centrifuged at 14,500 rpm (∼25,000 g) for 30 min at 4 °C. The pellet was then resuspended in binding buffer without 0.1% of BSA. Bradford protein assay (Bio- Rad, Hercules, CA) was used to determine the protein concentration and the membranes were diluted to 1000 μg/mL and stored in a −80 °C freezer for later use.

Radioligand binding assay experiments were conducted in 96-well polypropylene plates containing 5 μL of various concentrations of the test compounds diluted using 100% DMSO vehicle, 345 μL of binding buffer (50 mM Tris, 5 mM MgCl₂, 0.1% BSA pH 7.4), 50 μL of [³H]hGal (final concentration 2.0 nM; K_D = 3.71 nM; NOVANDI Chemistry AB, 93 Ci/mmol SA), and 100 μL of membranes (10 μg/well). All compound dilutions were tested in triplicate and the competition reaction started with the addition of the cell membranes; the plates were incubated for 150 min at 25 °C. Nonspecific binding was determined using 10 μM final concentration of M40. Incubation was terminated by rapid filtration through PerkinElmer Uni-Filter-96 GF/C presoaked in 0.5% polyethylenimine, using a Brandel 96-Well Plates Harvester manifold (Brandel Instruments, Gaithersburg, MD). The filters were washed a total of three times with 3 mL (3 × 1 mL/well) of ice-cold binding buffer. After drying, 65 μL PerkinElmer MicroScint20 Scintillation Cocktail was added to each filter well. Plates were counted using a PerkinElmer MicroBeta Microplate Counter. For each experiment, aliquots of the prepared radioligand solutions were measured to calculate the exact amount of radioactivity added, taking in account the experimentally determined top-counter efficiency for the radioligand. K_i values have been extrapolated by constraining the bottom of the dose–response curves (= 0% residual specific binding) in the nonlinear regression analysis. K_i values were calculated using GraphPad Prism 10 version 10.2.3 for Macintosh (GraphPad Software, San Diego, CA) utilizing One site- Fit K_i model. K_D value of the radioligand was determined separately via binding saturation experiments. K_i values were determined from at least three independent experiments performed in triplicate and are reported as mean ± SEM in three significant digits.

Kinetic measurements of tracer binding to HiBiT-GALR1

For binding kinetic measurements of six tracers to HiBiT-GALR1, the transfected HEK293 cells were seeded in a white 384-well plate (36 μL/well). 22-24 hours after transfection, 36 μL 2xHiBiT detection regent in Opti-MEM without phenol red, which consists of 100-fold dilution of LgBiT and 50-fold dilution of Nano-Glo HiBiT extracellular substrate, was added into each well. For background correction, a final concentration of 20 μM galanin peptides (prepared in Opti-MEM solution with 10% DMSO) was added to the control wells to determine nonspecific binding. After mixing gently for 15 min, 4 μL of serially diluted tracers (prepared as 20X solution in Opti-MEM solution with 20% DMSO and 30% tracer dilution buffer) were added to the cells. Kinetic reads were immediately measured on a PHERAstar FSX microplate reader.

Internalization assay

For internalization assay, pATG4092- HiBiT-GALR1 construct (0.05 μg) combined with transfection carrier DNA (9.95 μg) was transfected into 9 mL 2.2 x 10⁵ HEK293 cells. Transfected cells were seeded in a white 1536-well plate (5 μL/well) and incubated at 37 °C and 5% CO₂ for 22-24 hours. Serially diluted peptide ligands in DMSO solution were transferred to each well by Echo 650 acoustic liquid handler. The plate was then incubated at 37 °C and 5% CO₂ for 60 min. After cooling to room temperature, 5 μL 2xHiBiT detection regent in Opti-MEM without phenol red, which consists of 100-fold dilution of LgBiT and 50-fold dilution of Nano-Glo HiBiT extracellular substrate, was added into each well. Following 15 min of mixing at an orbital plate shaker, luminescence signal was measured on a PHERAstar FSX microplate reader.

Data analysis

All data were analyzed using GraphPad Prism software. For saturation binding, total binding, nonspecific binding, and corrected binding of each tracer were graphed with a log (inhibitor) versus response- variable slope fit. The saturation binding curve of each tracer was graphed using a one site-specific binding fit. For competitive displacement, normalization was carried out with respect to the no peptide neutral control, and the graph was done by using log (inhibitor) versus response- variable slope fit. The derived IC₅₀ values were utilized to calculate K_i values using the Cheng–Prusoff equation: K_i = IC₅₀/(1+ [L]/K_D). where [L] is the concentration of a tracer in the assay, and K_D is its affinity in a saturation binding experiment. For kinetic analyses, data of corrected BRET values of serially diluted tracers were graphed using the association kinetics: two or more concentrations of hot fit. The kinetic constants (k_on and k_off) and binding constant (K_D) of the tracer were determined from the curves. The residence time (τ) was determined from the k_off: τ = 1/k_off. For internalization experiments, data were normalized to a no peptide neutral control and graphed with a log (inhibitor) versus response-variable slope fit.

Results:

Design, syntheses, and characterization of peptide tracers.

NanoBRET target engagement assays require a tracer - an appropriate ligand modified with a fluorescent group - that can serve as an energy acceptor for bioluminescence resonance energy transfer (BRET) from the HiBiT/LgBiT NanoLuc luciferase donor complex incorporated into the target protein. As there are no known small molecules that bind to GALR1, we conceived a BRET peptide tracer design based on its cognate endogenous ligand galanin. Fortunately, galanin contains a single naturally occurring lysine residue (K25) in its sequence, and the primary amine of the K25 side chain can be employed as a handle for incorporation of a chosen fluorophore acceptor (Figure 3). In addition to galanin, truncated galanin 1-15 (Gal(1-15)) and the 20-amino acid ligand M40 were also chosen as candidate peptides for tracer design. Gal(1-15) was chosen because it has all the minimum number of critical amino acids required for agonist activity and would be easier to synthesize than longer derivatives.^{12, 15} This design was also appealing because it has been shown that while the N-terminus residues (G1 – V16) of endogenous galanin peptide are critical for receptor binding, the C-terminus can tolerate truncation or modification.¹⁶ M40 was chosen as it was initially described as a potent antagonist,¹⁷ although more recently it has been shown to be a potent and efficacious agonist.¹⁸ As naturally occurring lysine was not present in these two peptides, an extra lysine was incorporated at the C-terminal of each peptide; thus, Gal(1-15+K) and M40(+K) were synthesized.

Figure 3.

Chemical structures of all six BRET tracers and their activity in the PathHunterβ -arrestin assay. A. Chemical structures of six BRET tracers. B. Function of six tracers evaluated by PathHunter GALR1 β-arrestin assay. Briefly, stable cell line expressing GALR1 β-arrestin was seeded in 1536-well plate and cultured at 37°C 5% CO₂ overnight. Increasing Cells were treated with increased concentrations of tracers were added to the cells and incubated at 37°C 5% CO₂ for 90 min. PathHunter^® detection reagent was added to cells and luminescent signal was detected on ViewLux. C. Imaging of binding and displacement of Tracer 1 to HiBiT-GALR1 transiently expressed on HeLa cells. Panels are pseudocolored for the donor (blue) and acceptor (red) channel.

The naturally occurring lysine residue in galanin peptide and incorporated lysine at the C-termini of peptides Gal(1-15+K) and M40(+K) provided a bioconjugation handle for coupling the BODIPY-based dye, NanoBRET^™ 590. The reactive BODIPY dye reagents contain N-hydroxysuccinimide esters (NHS) for amine conjugation and are commercially available as NanoBRET^™ 590 SE Dye and NanoBRET^™ 590-C4-SE Dye, which differ only in the carbon chain linker length between the BODIPY fluorophore and NHS.¹³ As BRET efficiency depends on both proximity (<10 nm) and geometry between the donor and acceptor,¹¹ a shorter and a longer linker length were deliberately incorporated into the tracer design. This would allow us to investigate which of our newly synthesized tracer constructs would bind to GALR1 while also presenting the acceptor fluorophore (NanoBRET^™ 590) within the optimal distance and orientation for efficient energy transfer between the luminescence donor (HiBiT/LgBiT) and NanoBRET^™ 590.

To generate tracer peptide constructs 1-6 (Figure 3A), a modified synthetic strategy was employed (Figure 2). Starting with Fmoc protected lysine, with a Boc-protected side chain amine, amino acids were sequentially attached with automated solid phase peptide synthesis to a Rink amide MBHA resin on a CEM Liberty Blue^™ instrument. Instead of a final Fmoc cleavage of the N-terminal glycine, the peptides were first cleaved from their solid support by suspension with a trifluoroacetic acid deprotection cocktail that removed all the protecting groups from the side chains and generated a primary amide at the C-terminus lysine. These conditions also cleaved the Boc protecting group on the side chain of the lysine amino acid designated for attachment of our NanoBRET^™ 590 Dye, while the N-terminal peptide amine remained Fmoc protected and unreactive. The crude resin-cleaved Fmoc-protected peptides were precipitated by the addition of cold diethyl ether and purified by preparative RP-HPLC. Following this, galanin, Gal (1-15+K) and M40(+K) peptides were conjugated with NanoBRET^™ 590 dyes before being subjected to Fmoc deprotection and immediately purified by RP-HPLC. The peptide tracers were characterized by HPLC and high-resolution mass spectrometry and stored as solid powders at −20 °C. The spectral properties of the BODIPY fluorophore in the tracers remained unchanged after conjugation (Supplemental Figure S1). 10 mM solutions in DMSO were prepared for use in biological assays; we found slight decomposition, as observed by LCMS, for these DMSO solutions that had been stored for > 60 days (data not shown).

Figure 2.

Representative synthesis of peptide tracers, e.g. Gal(1-15+K)-NanoBRET^®590. Solid phase peptide synthesis was performed on a CEM Liberty Blue^™ automated microwave peptide synthesizer. Fmoc protected N(Boc) lysine was coupled to Rink amide MBHA resin with ethyl(hydroxyimino)cyanoacetate DIC (Oxyma)/ N,N′-diisopropylcarbodiimide (DIC) followed by Fmoc deprotection using 20% piperidine. Subsequent amino acids were coupled using the same procedure. Fmoc protected Gal(1-15+K) attached to the resin was cleaved using trifluoroacetic acid (TFA), 5% H₂O, 5% Phenol, and 2% triisopropyl silane (TIPS). The Fmoc protected-Galanin peptide precursor was coupled to NanoBRET^™ 590 SE Dye with DIPEA in DMF followed by addition of 20% piperidine to remove the Fmoc at the N-terminal glycine.

To confirm that the peptide tracers (Figure 3A) retain the ability to bind and activate GALR1, we performed the PathHunter β-arrestin recruitment assay (DiscoveRx/Eurofin). In this functional assay, agonist-induced receptor activation and subsequent recruitment of β-arrestin to the receptor leads to the fragment complementation of β-galactosidase, which can be monitored by chemiluminescent signal from the turnover of the fluorogenic 4-MU-β-gal substrate ¹⁹. After stimulating with increased concentrations of tracers, concentration-response curves were obtained for all tracers, which indicated that they can bind and activate GALR1 (Figure 3B). Tracers 1 and 2, derived from Gal1-15 showed similar potency (tracer 1, EC₅₀ = 14.9 ± 0.9 nM; tracer 2, EC₅₀ = 11.3 ± 1.1 nM). Tracers 3 and 4, derived from M40 peptide, had the most potent β-arrestin activity with EC₅₀s of 5.9 ± 0.5 nM and 6.4 ± 0.4 nM respectively. Tracer 5, derived from galanin peptide with a short linker, showed the weakest potency (EC₅₀ = 33 ± 1.4 nM). However, tracer 6, which was also derived from galanin peptide but had the extended linker, showed better potency (EC₅₀ = 8.7 ± 0.2 nM).

To visualize the specific engagement of tracers to the receptor, representative tracer 1 was incubated with HeLa cells overexpressing HiBiT-GALR1 in the presence or absence of galanin. Filtered bioluminescence images were taken to visualize the specific engagement of tracer 1 with HiBiT-GALR1 transiently expressed on the surface of these living cells. Complementation with LgBiT, followed by treatment with tracer 1 (1 μM) and substrate, resulted in efficient BRET (observed by fluorescent signal output), which was significantly diminished through competitive displacement by 10 μM (unlabeled) galanin (Figure 3C). This attenuation of BRET confirmed specific and reversible interaction between tracer 1 and HiBiT-GALR1. The ability to carry out such analyses with transiently expressed HiBiT-GALR1 in live cells using agonist ligands demonstrates the sensitivity of this approach. Furthermore, signal localized to the cell surface further validated that signal originated from HiBiT-tagged receptors localized to cell-surface.

Saturation binding and competitive binding of peptide tracers.

To evaluate the saturation binding of all six tracers to GALR1, HEK293 cells were transiently transfected with HiBiT-GALR1 plasmid. Cells were treated with increasing concentrations of tracers for 90 min in the absence (total BRET) and presence of 33 μM unmodified galanin peptide (nonspecific BRET control) (Figures 4 A, D, G, J, M and P). As the tracers showed agonist activity in the GALR1 β-arrestin assay (Figure 3B) the experiments were performed at room temperature to minimize agonist-induced receptor internalization which was observed at 37 °C. At 33 μM, unmodified galanin, was able to completely suppress the BRET signal at tracer concentrations <1 μM for all six tracers; we observed the same with 30 μM of M40 which could also be used to determine nonspecific BRET (Supplemental Figure S3). Specific BRET was determined by subtracting nonspecific BRET values from total BRET values (Figures 4 B, E, H, K, N and Q). The total BRET, nonspecific BRET and specific BRET of each tracer were plotted against the tracer concentration on a log scale and fitted with a sigmoidal curve. Specific BRET of each tracer was also plotted versus tracer concentrations and fitted with one site-specific binding fit to determine the K_D value of each tracer (Figure 4 C, F, I, L, O, and R).

Figure 4.

BRET saturation binding curves of tracers 1-6 to GALR1. HEK293 cells expressing HiBiT-GALR1 were used to evaluate the saturation binding of tracers to the GALR1. 24 hours after transfection of HiBiT-GALR1 plasmid, cells were incubated with increased concentrations of tracers in the absence and presence of 33 μM galanin peptides (A, D, G, J, M and P). After equilibrium binding at room temperature for 90 min, Nano-Glo^® HiBiT extracellular detection reagent was added, and BRET signal was measured on a PHERAstar FSX microplate reader equipped with 450/80BP and 610/LP filters. The specific binding curves represent the difference between total binding and nonspecific binding (in the presence of 33 μM galanin) (B, E, H, K, N and Q). Saturation binding curves were plotted in GraphPad Prism (C, F, I, L, O, and R). Each experiment was performed in triplicate (n = 3).

All six tracers showed a clear concentration-dependent and saturable increase of BRET signals (total BRET). Notably, BRET signal intensities were negatively correlated with the length of the peptide tracers. Highest BRET signals were observed for tracers 1 and 2 (Gal (1-15) derived tracers, 16 aa), moderate BRET for tracers 3 and 4 (M40 derived tracers, 21 aa), and lowest BRET for tracer 5 and 6 (galanin derived tracers, 30 aa) (Figure 4 A, D, G, J, M and P). This suggests that the shorter Gal(1-15) peptide tracer can adopt a conformation favorable for efficient energy transfer from the donor (HiBiT/LgBiT) to the acceptor (BODIPY fluorophore). In the presence of excess unmodified galanin, nonspecific BRET was generally low for tracers 1, 2 and 5 (Figures 4 A, D, M) but higher for tracers 3, 4 and 6, especially ≥1 μM (Figures 4 G, J, P). While nonspecific BRET can be an issue depending on what concentration between EC₅₀ to EC₈₀ is chosen for the BRET assay, the affinity of a tracer to its receptor can also play a crucial role in its utility as a BRET tracer. To that end, we obtained a range of binding affinities for the six tracers as represented by the calculated equilibrium dissociation constants (K_D) (Table 1). M40 derived tracers 3 and 4 showed high binding affinity (tracer 3, K_D = 10.0 ± 0.3 nM; tracer 4, K_D = 11.4 ± 0.4 nM). Galanin derived tracers showed variable binding affinities (tracer 5, K_D = 126.4 ± 2.0 nM; tracer 6, K_D = 19.3 ± 1.8 nM). Gal (1-15) derived tracers showed moderate binding affinity (tracer 1, K_D = 183.1 ± 16.9 nM; tracer 2, K_D = 192.0 ± 7.3 nM) (Figure 4 C, F, I, L, O, and R; Table 1). It is interesting to note that the linker length did not have much of an effect on the binding affinities of tracers derived from Gal(1-15) and M40 peptides, however, it significantly affected the binding affinities of tracers derived from galanin (tracer 5, K_D = 126.4 ± 2.0 nM vs. tracer 6, K_D = 19.3 ± 1.8 nM).

Table 1.

Binding affinity of tracers in both equilibrium and real time (Mean ± SEM, n=3)

Tracer	Name	Equilibrium	Kinetic
		KD (nM)	KD (nM)	kon (M−1min−1)	koff (min−1)	τ (min)
1	Gal(1-15+K)-NanoBRET®590	183.1 ± 16.9	40.0 ± 6.8	7.1 ± 0.01 x105	0.023 ± 0.003	46.2 ± 6.8
2	Gal(1-15+K)-C4-NanoBRET®590	192.0 ± 7.3	105.2 ± 4.5	6.7 ± 0.02 x105	0.071 ± 0.004	14.3 ± 1.2
3	M40(+K)-NanoBRET®590	10.0 ± 0.3	10.2 ± 2.8	5.9 ± 0.4 x106	0.063 ± 0.021	20.5 ± 6.8
4	M40(+K)-C4-NanoBRET®590	11.4 ± 0.4	18.0 ± 1.8	2.4 ± 0.06 x 106	0.043 ± 0.003	23.5 ± 1.7
5	Galanin-NanoBRET®590	126.4 ± 2.0	26.9 ± 7.3	3.8 ± 0.9 x 106	0.090 ± 0.004	11.1 ± 0.5
6	Galanin-C4-NanoBRET®590	19.3 ± 1.8	38.6 ± 3.1	1.7 ± 0.2 x 106	0.063 ± 0.003	16.0 ± 0.7

To further evaluate the binding of tracers to GALR2 and GALR3, HEK293 cells were transiently transfected with HiBiT-GALR2 and HiBiT-GALR3 plasmids. Cells were treated with increasing concentrations of tracers for 90 min in the absence (total BRET) and presence of 33 μM unmodified galanin peptide (nonspecific BRET). A clear dose-dependent and saturable increase of BRET signals was observed in HEK293 cells expressing HiBiT-GALR2 (Supplemental Figure S4). BRET signal intensities were negatively correlated with the length of the peptide tracers, which was similar to that of GALR1. The equilibrium K_D determined for the tracers in the HiBiT-GALR2 BRET assay (Supplemental Table 2) was comparable to what was obtained with HiBiT-GALR1 (Table 1); their GALR2 β-arrestin activities (Supplemental Table 3) were also comparable to their GALR1 counterparts (Figure 3B). However, saturable BRET signals were not observed in HEK293 cells expressing HiBiT-GALR3 (Supplemental Figure S5), which might indicate poor cell-surface expression of GALR3 as reported previously.²⁰

To further evaluate the suitability of these fluorescent tracers, we synthesized and analyzed a panel of peptides derived from galanin (Supplemental Table 1) from the literature with a range of affinities as determined previously in radioligand displacement assays.^{3, 21} To examine competitive tracer displacement, HEK293 cells expressing HiBiT-GALR1 were treated with increasing concentrations of these peptide ligands in the presence of a fixed EC₈₀ concentration of the tracers (Figures 5 A-F). A dose-dependent decrease of BRET signal was consistently observed for seven out of ten ligands regardless of which tracer was used. The K_i value of each unmodified galanin peptide was calculated using the Cheng–Prusoff equation. These peptide ligands can be classified into four groups according to their binding affinities (Figures 5 A–F and Table 2): high binding affinity (K_i < 0.2 μM for Galanin, M15 and M35), moderate binding affinity (K_i = 0.2 μM – 2.0 μM for M40, M617 and Gal(1-15)), low binding affinity (K_i = 2 μM – 20 μM for Gal(DMG-(2-15)) and inactive (K_i > 20 μM for Gal (1-11), Gal (2-11), and M871)). The rank order of the affinities of these peptides broadly agreed with their reported K_i values in radiolabeled binding assays (Supplemental Table 1).

Figure 5.

Characterization of binding affinities for unmodified galanin peptide and 9 peptide derivatives using tracer displacement assay. Briefly, 24 hours after transfection of HiBiT-GalR1 plasmid, cells were incubated for 10 min with increased concentrations of unmodified galanin peptide and 9 derivatives. ~EC₈₀ concentration of tracers were then added to the cells, which were further incubated at room temperature for 90 min. Nano-Glo^® HiBiT extracellular detection reagent was added, and BRET signal was measured on a PHERAstar FSX microplate reader equipped with 450/80BP and 610/LP filters. BRET values were normalized to the maximal BRET value in each tracer. Each experiment was performed in triplicate (n = 3). G The functional activity of these unmodified peptides in the PathHunter β-arrestin recruitment assay. H The correlation between the binding affinity (pK_i) and the activity of unmodified galanin derived peptides (pEC₅₀). The pK_i was calculated from K_i values (pK_i = − log[K_i]), which was derived from NanoBRET^™ assays using the six tracers and calculated according to the Cheng–Prusoff equation, and pEC₅₀ was calculated from EC₅₀ values (pEC₅₀ = −log[EC₅₀]), which were derived from PathHunter β-arrestin assay. Correlation was determined by linear regression analysis using GraphPad Prism 10 software.

Table 2.

K_i values (Mean ± SEM, n=3) of unmodified peptides calculated from NanoBRET assay.

	Tracer 1 (nM)	Tracer 2 (nM)	Tracer 3 (nM)	Tracer 4 (nM)	Tracer 5 (nM)	Tracer 6 (nM)
Galanin	116.2 ± 6.1	69.1 ± 0.8	94.6 ± 3.3	63 ± 3.3	125.1 ± 7.4	41.6 ± 3.3
M15	33.7 ± 0.6	17.8 ± 0.9	24.7 ± 1.6	20.7 ± 1.1	36.7 ± 1.8	17.7 ± 1
M35	53.1 ± 1.2	27.3 ± 1	46.9 ± 2.4	34.9 ± 3.1	59 ± 3.4	26.7 ± 2.9
M40	957.1 ± 70.7	589 ± 25.5	632.3 ± 38.9	599.1 ± 64.1	796.7 ± 25.8	242.7 ± 32.5
M617	678.6 ± 37.9	458.3 ± 23.6	616.1 ± 45.4	434.6 ± 33.7	1112.7 ± 200.5	345.2 ± 18.7
M871	>20000	>20000	>20000	>20000	>20000	>20000
Gal(1-15)	1070 ± 60.3	699 ± 24.2	1001.1 ± 80.5	622.6 ± 51.3	1093.7 ± 47.7	649.1 ± 72.4
Gal(DMG,2-15)	4056.6 ± 455	11488.4 ± 2643.9	8650.6 ± 768	5125.7 ± 663.6	7935.6 ± 1356.5	19862.5 ± 10166.1
Gal(1-11)	>20000	>20000	>20000	>20000	>20000	>20000
Gal(2-11)	>20000	>20000	>20000	>20000	>20000	>20000

Galanin, M15 and M35 have <1 nM K_i values reported in radiolabeled galanin displacement assays (Supplement Table 1) and were the most potent in the BRET assays.³ However, there was a considerable drop (~50-500 fold) in the potencies in the cell-based BRET assay (Table 2). We grouped with M617, Gal(1-15) and M40, the peptides that formed the basis for tracers 3-6 ²². These three peptides also have potent reported K_i values < 3 nM in radiolabeled binding assays (Supplement Table 1) but their BRET K_i values are blunted, although their normalized BRET signal is almost completely subdued at the highest concentration tested. These observations emphasize how the BRET assay can pick apart nuances in activities between different ligands. M871 and Gal (2-11), which don’t contain the first critical N-terminal glycine, showed no decrease of BRET signals even at highest concentration (20 uM) used in the experiment (Figures 5 A–F). This agrees with their reported K_i values of 420 and >5000 uM respectively against human GALR1 in radiolabeled displacement assays (Supplemental Table 1). A derivative of Gal(1-15) where the glycine crucial for activity was replaced with an N,N-dimethylglycine (DMG) was synthesized (Gal(DMG,2-15)) to check the effect of a dimethyl amine replacement of the primary amine. This peptide was able to show only a partial response that was <50% even at the highest tested concentration; we observed the same with Gal(1-11).²³

The activities of these peptides were further evaluated in the PathHunter β-arrestin recruitment assay. They showed a range of activities in this functional assay (Figure 5G and Table 3), six peptides showed high potency (3 – 10 nM) in activating GALR1 (M15 > M35 > M40 > galanin > M617 > Gal (1-15)), while four peptides showed low potency (Gal(DMG,2-15) > Gal(1-11) > M871 > > Gal (2-11)). To evaluate the correlation between BRET binding data and β-arrestin recruitment assay data, we performed linear regression analysis between the pK_i and pEC₅₀ values in the respective assays. As the K_i values of Gal (1-11), M871, and Gal (2-11) were > 20 μM and deemed inactive, they were removed from the analysis. Our analysis found that there is a correlation between the BRET pK_i values (binding affinity), calculated from each tracer, and the β-arrestin pEC₅₀ values (functional activity) (Figure 5 H).

Table 3.

Potency of unmodified peptides in PathHunter GALR1 β-arrestin assay.

Peptides	EC50 (nM) Mean ± SEM (n=3)
Galanin	6.3 ± 0.2
M15	3.2 ± 0.4
M35	3.7 ± 0.4
M40	6.2 ± 0.5
M617	8.3 ± 0.8
M871	615 ± 32
Gal(1-11)	100.8 ± 8.8
Gal(2-11)	15,667.5 ± 2,494.1
Gal(1-15)	10 ± 1.2
Gal(DMG,2-15)	46 ± 2.7

Radioligand binding assay of fluorescently labeled peptide tracers.

To compare our findings in the BRET assay, five of the six peptide tracers were evaluated by competitive radioligand binding assay using a fixed concentration of [³H] human galanin competing with a series of concentrations of the peptide tracers. In this experiment, membrane preparations from HEK293 cells stably expressing GALR1 were employed.

The binding affinities of the peptide tracers in the radiolabeled assay followed a trend that was similar to the K_D values obtained in the BRET assay (Table 4). M40 derived tracers 3 and 4 showed the highest binding affinities among the tested compounds (tracer 3, K_i = 2.10 ± 0.269 nM; tracer 4, K_i = 1.98 ± 0.337 nM). Galanin derived tracer 5 (K_i = 9.79 ± 0.383 nM) and Gal (1-15) tracer 2 (K_i = 9.87 ± 1.84 nM) had similar binding affinities. The Gal (1-15) derived tracer 1 showed the lowest binding affinity of all the tested compounds (K_i = 33.8 ± 3.23 nM; Table 2). Additionally, both M40 derived tracers (tracer 3, K_i = 2.10 ± 0.269 nM; tracer 4, K_i = 1.98 ± 0.337 nM) demonstrated a small increase in affinity when compared to the parent peptide M40 (K_i = 10.2 ± 1.65 nM) indicating that the incorporation of lysine at the C-terminus of M40 and the addition of the BODIPY dye was well tolerated. Although the addition of BODIPY in K25 of Galanin derived tracer 5 (K_i = 9.79 ± 0.383 nM) decreased the binding affinity when compared to the parent peptide, Galanin (K_i = 2.10 ± 0.378 nM), this reduction is only ~5 fold, which is a remarkably low compared to other small molecule fluorescently-labelled ligands reported in the literature.^{24, 25} Moreover, when we compared tracer 5 with galanin in a cAMP functional assay, we did not see any discernable difference in activity (Supplemental Figure S2) indicating the HiBiT tag or fluorophore did not interfere with function in cellular assays.

Table 4.

K_i values of unmodified peptides and tracers calculated from radioligand binding assay.

Tracer	Name	Ki (nM) Mean ± SEM (n=3)
	Galanin	2.10 ± 0.378 (6)
	M40	10.2 ± 1.65 (3)
1	Gal(1-15+K)-NanoBRET®590	33.8 ± 3.23 (3)
2	Gal(1-15+K)-C4-NanoBRET®590	9.87 ± 1.84 (3)
3	M40(+K)-NanoBRET®590	2.10 ± 0.269 (3)
4	M40(+K)-C4-NanoBRET®590	1.98 ± 0.337 (3)
5	Galanin-NanoBRET®590	9.79 ± 0.383 (3)

To evaluate the correlation of binding affinities of five tracers calculated from radioligand binding assay and HiBiT based ligand binding assay, the pK_i calculated from radioligand binding assay and the pK_D calculated from HiBiT based ligand binding assay was tested by linear regression analysis in GraphPad Prism 10 software. There is a correlation (R²=0.845) between the two binding assays (Supplemental figure 6).

Binding kinetics of peptide tracers.

To determine binding kinetic parameters of each NanoBRET peptide tracer, HEK293 cells expressing HiBiT-GALR1 were treated with LgBiT and furimazine for 15 min to allow the formation of HiBiT/LgBiT complex and production of bioluminescence. Increased concentrations of tracers were added to the cells and BRET signals were continuously measured over time. The specific binding of peptide tracers was calculated by subtracting BRET values in the presence of excess unmodified galanin peptides from BRET values in the absence of galanin peptides at each time point (Figure 6 and Table 1). The kinetic parameters calculated from association kinetic curves showed that the association rates (k_on) of all tracers were within 10-fold of range: tracer 3 was fastest (5.9 ± 0.4 x10⁶ M⁻¹min⁻¹) and tracer 2 (6.7 ± 0.02 x10⁵ M⁻¹min⁻¹) was slowest. The dissociation rates (k_off) of all tracers were within four-fold range: tracer 1 was slowest (0.023 ± 0.003 min⁻¹) and tracer 5 was fastest (0.090 ± 0.004 min⁻¹). Therefore, tracer 1 had the longest residence time (46.2 ± 6.8 min) and tracer 5 had the shortest residence time (11.1 ± 0.5 min). The K_D values calculated from the association kinetic curves were comparable to those calculated from the saturation binding except for tracers 1 and 5 whose kinetic K_D values were 4-5-fold lower than those calculated from saturation binding respectively.

Figure 6.

Association kinetics of 6 galanin tracers. HEK293 cells expressing HiBiT-GALR1 were used to evaluate the association kinetics of tracers to GALR1. 24 hours after transfection of HiBiT-GALR1 plasmid, Nano-Glo^® HiBiT extracellular detection reagent was added to cells and incubated at room temperature for 15 min to form HiBiT/LgBiT luciferase complex. Increased concentrations of tracers in the absence and presence of 33 μM galanin peptides were added to the cells, and BRET signal was continuously measured over 30 min on a PHERAstar FSX microplate reader equipped with 450/80BP and 610/LP filters. Specific BRET signal was determined by subtracting BRET values in the presence of 33 μM galanin peptides from BRET values in the absence of galanin peptides. Each experiment was performed in triplicate (n = 3).

Ligand-induced receptor internalization.

HiBiT-GALR1 localized on the plasma membrane is expected to be internalized by clathrin-dependent endocytic pathways upon stimulation with agonists. The percentage of internalized receptor can serve as an indirect indicator of plasma membrane GALR1 activation. To test this approach, cells expressing HiBiT-GALR1 were treated with galanin at 37 °C for 60 min and monitored for dose dependent agonist-induced receptor internalization. As LgBiT is not cell permeable, it can only form the HiBiT/LgBiT complex with HiBiT-GARL1 receptors remaining on the cell surface. Therefore, decreased luminescence signal, on addition of LgBiT and substrate (furimazine), will represent receptor loss at the cell surface caused by ligand-induced internalization (Figure 7A). To maximize the assay detection window, different transfection protocols with a range of plasmid concentrations were evaluated. We found that galanin induced dose-dependent receptor internalization at 1 x 10⁻² and 1 x 10⁻³ μg/mL, but not at 1 x 10⁻¹ μg/mL plasmid transfection (Figure 7B). Thus, the percentage of internalization was inversely related to the expression level of receptors (Figure 7B). We also verified that receptor internalization did not occur at room temperature (25 °C ) or at 4 °C (Supplemental Figure S7). 1 x 10⁻³ μg/mL of plasmid transfection gave us the maximum detection window and was used in the internalization assay to evaluate the same set of peptide ligands that had been evaluated in the BRET target engagement and PathHunter β-arrestin recruitment assays. To that end, increasing concentrations of peptide ligands (0.34 nM – 2.2 μM) were incubated with HEK293 cells expressing HiBiT-GALR1 receptor. A concentration-dependent decrease of luminescence signal was observed for eight out of ten ligands (Figure 7C). The binding affinities of the peptides were shown in Table 5. M871 and Gal (2-11) showed no decrease of luminescence signal even at the highest assay concentration of 2.2 μM used in the experiment (Figure 7C). We also tested these two peptides at higher concentrations (20 μM) and found that M871 showed moderate activity (~10%) while Gal (2-11) showed no activity (data now shown). M15 (IC₅₀ = 5.3 ± 1.0 nM) and M35 (IC₅₀ = 4.0 ± 0.4 nM) showed high ligand-induced receptor internalization. It was interesting that M40 showed partial agonist activity (efficacy ~50%) in the internalization assay (Figure 7C), which was different from the full agonist activity we observed in the PathHunter β-arrestin recruitment assay (Figure 5G) and cAMP assay (data not shown). M40 has been reported as an antagonist of the GALR1 and GALR2 in animal studies.^{26, 27} The assay-dependent activity of M40 in our studies indicates that M40 might be a functionally selective ligand, and the nature of its GALR1 modulation is likely dependent on the context of the assay being used to evaluate its activity. Thus, its differential behavior should be teased out by analyzing it in different modes. To evaluate the correlation of activities of peptides between internalization assay and PathHunter β-arrestin recruitment assay, we performed linear regression analysis of the data. As the IC₅₀ of two peptides (Gal (2-11), M871) were >20 μM in the internalization assay, they were removed from the analysis. Our analysis found that there is a correlation (R²=0.739) between pIC₅₀ value of internalization activity and pEC₅₀ value of PathHunter β-arrestin recruitment assay (Figure 7D).

Figure 7.

Characterization of activity of unmodified galanin peptide and 9 derivatives using HiBiT-GALR1 internalization assay. A. Schematic of HiBiT-GALR1 internalization assay. In this assay, HiBiT tag is attached to the N-terminus of GALR1. Agonist peptides can induce the internalization of HiBiT-GALR1, which decreases the number of receptors and HiBiT at the cell surface. Consequently, fewer HiBiT/LgBiT luciferase complexes will form, resulting in a reduced bioluminescent signal upon the addition of the NanoLuc luciferase substrate. B. The effect of transient expression level of GALR1 in HEK293 cells on galanin induced internalization. 0.001, 0.01, 0.1 μg/mL HiBiT-GALR1 plasmid were transfected into HEK293 cells for 24 hours and galanin induced internalization of receptors was measured after cells were stimulated for 1 hour by increasing concentrations of galanin at 37 °C and 5% CO₂. C. 0.005 μg/mL HiBiT-GALR1 plasmid were transfected into HEK293 cells for 24 hours. Cells were treated with increasing concentrations of galanin and 9 derivative peptides and incubated at 37 °C 5% CO₂ for 1 hour. Detection regents were added and the activity of each peptide was calculated by decreased luminescence output. Each experiment was performed in triplicate (n = 3). D. The correlation between activities of unmodified galanin peptide and 9 derivatives in HiBiT-GALR1 internalization assay and PathHunter β-arrestin recruitment assay. The pIC₅₀ was calculated from IC₅₀ value (pEC₅₀ = −log(IC₅₀)) of peptides in the HiBiT-GALR1 internalization assay, and the pEC₅₀ was calculated from EC₅₀ values (pEC₅₀ = −log(EC₅₀)) of peptides in PathHunter β-arrestin assay. Correlation was determined by linear regression analysis using GraphPad Prism 10 software. NanoLuciferase signal was a PHERAstar FSX microplate reader equipped with a luminescence filter pair (450 nm bandpass filter and 610 nm longpass filter). The focal height is 12.5mm and the gain setting is 3600.

Table 5.

Potency of unmodified peptides in GALR1 internalization assay (curves in Figure 7C).

Peptides	IC50 (nM) Mean ± SEM (n=3)
Galanin	36.7 ± 8.9
M15	5.3 ± 1
M35	4 ± 0.4
M40	534.3 ± 221.1
M617	77.2 ± 3.6
M871	>20000
Gal(1-15)	93.5 ± 9.5
Gal(DMG,2-15)	6363.7 ± 848.1
Gal(1-11)	1917 ± 66.8
Gal(2-11)	>20000

Discussion:

While radioligand binding assays have been widely used in pharmacological research, in recent years the cell-based NanoBRET^™ assay has emerged as an advantageous alternative to evaluate ligand binding across a wide range of GPCR families in live cells ^{5–8, 28–31}. In earlier studies, existing NanoBRET tracers were often derived from small molecule ligands, the development of peptide-based tracers remains underexplored, particularly for neuropeptide receptors. Our rational design of fluorescently labeled peptide tracers tailored to GALR1 represents a significant innovation, as it maintains high affinity and specificity while enabling robust detection of GALR1 in live cells. The high affinity of peptide-based tracers to the receptors is particularly advantageous for resolving low-abundance receptor populations, where other traditional techniques may lack the necessary sensitivity. We demonstrate that peptide-based tracers can be successfully optimized for neuropeptide GPCRs and open new avenues for studying these therapeutic targets with high precision in native-like conditions.

Compared to the radioligand displacement assay, there are several advantages of this cell-based NanoBRET assay. The NanoBRET^™ assay is performed in live cells and thus ligand binding to the receptors is expected to yield more physiologically relevant binding parameters. On the other hand, in radioligand binding assays, cells are lysed, plasma membranes are washed and suspended in a Tris buffer. This preparation process removes vital intracellular components such as G-proteins, β-arrestins, and ions which may allosterically modulate receptors and thereby influence the binding of ligands ^{32, 33}. For example, sodium stabilizes the inactive state of many class A GPCRs by interacting with a highly conserved aspartate residue Asp^2.50 in an allosteric site,^{34, 35} which exists in GALR1. Zinc ions can also allosterically modulate the signaling of GPCRs ³⁶; a cryo-EM structure and follow-up pharmacological studies have shown that zinc ion is a negative allosteric regulator of GALR1.³⁷ In addition to intracellular components, components in the plasma membrane such as lipids,^{34, 38, 39} receptor activity-modifying proteins (RAMPs),⁴⁰ and other GPCRs,⁴¹ can also allosterically modulate signaling of GPCRs. Previous studies have shown that GALR1 can form heteromers with μ-opioid receptors (MORs) and promote the attenuation of MOR signaling.^{42, 43} Thus, membrane preparations might compromise the integrity and composition of the plasma membrane reducing the physiological context of radioligand binding assays.

Unlike radioligand binding assays, cell-based NanoBRET^™ assays allow us to monitor the real-time interaction of ligands with their receptors within live cells. The interaction of ligands and receptors is dynamic and undergoes constant conformational changes during extracellular association and dissociation of ligands, G-protein coupling and uncoupling, phosphorylation of receptors, and β-arrestin recruitment.⁴⁴ Finally, NanoBRET assays typically have lower experimental complexity compared to radioligand binding assays. The experiments can be performed in a standard laboratory setting, eliminating the need for specialized equipment and facilities required for radioactive measurement. The simplified assay setup, low experimental costs, and streamlined workflow of the NanoBRET^™ assays make them more accessible to scientists across different disciplines.

The first step in our consideration of this NanoBRET^™assay was tracer design. As we could not predict the optimal distance and overall conformation favorable for efficient energy transfer between the HiBiT/LgBiT donor at the N terminus of receptors and acceptor fluorophore (NanoBRET^™ 590) in the tracers, three galanin peptides with 15aa, 20aa and 30aa, and two linker lengths (short and long) were deliberately chosen to make six different tracers. To our surprise, all the tracers worked well in the assay. However, we noticed a negative correlation between the length of peptide and linker and the intensity of the BRET signal. Tracer 1, which has the shortest peptide length (15aa) and short linker, had the highest BRET signal, which suggests that it was able to adopt the most favorable conformation for efficient energy transfer between HiBiT/LgBiT and the BODIPY acceptor fluorophore. Tracer 6, which has the longest peptide length (30aa) and long linker, had the lowest BRET signal. However, the binding affinities of the tracers, which were determined by the dissociation constant (K_D), did not correlate to their BRET signal window. The K_D values of the tracers were calculated from two analyses in our experiments, one from saturation binding analysis in equilibrium, and another from kinetic analysis in real time. K_D values calculated from both analyses were closely matched to each other. Tracer 3 and 4 (derived from M40 peptide, 21aa length) had the highest binding affinity, followed by tracers 6 and 4 (derived from galanin, 30aa length), and tracer 1 and 2 (derived from Gal (1-15), 16 aa length). The binding affinities of tracers to the receptors correlated well with the β-arrestin recruitment functional assay; this established good correlation between cell-based binding and function. The BRET tracer binding affinities calculated from the tracer displacement assay were right shifted compared to their K_i values calculated from radioligand binding assay. Such right shifts are expected when determining ligand binding in cells compared to membrane preparations, as additional physiological factors in live-cell assays, such as membrane barriers and regulatory proteins may weaken ligand binding.

To validate our assay, we tested a panel of previously reported galanin derived peptides with a range of binding affinities for GALR1 in the cell based NanoBRET^™ assays. The rank order of the binding affinities of unmodified peptides was consistent for all tracers and broadly agreed with previously reported K_i values in radiolabeled displacement assays.^{3, 21} Importantly, the ranking of their BRET binding affinities correlated well with their potencies in our functional β-arrestin recruitment and internalization assays demonstrating a good correlation between binding and function between cell-based assay formats. However, we observed a significant right shift in the concentration-response curves with more than 10-fold K_i values compared to the K_i reported with radioligand displacement assay. The discrepancies in the binding affinities of these unmodified peptides obtained with our cell-based NanoBRET^™ assay versus the radioligand displacement assay underscore the significance of the cellular context, as discussed earlier. A potential factor could be the buffer; while a Tris buffer is utilized in the radioligand binding assay, our cell based NanoBRET^™ assays employed an assay medium containing fetal bovine serum (FBS). Albumin, the most prevalent protein in FBS, is recognized to bind to a variety of peptides.⁴⁵ FBS also contains peptidases which have been shown to decrease the stability of certain peptides.^{46, 47} Thus, the buffer may play a role in decreasing the bioavailability of galanin-derived peptide in the BRET assay. Such a discrepancy between cellular and acellular assays has also been observed with kinase BRET assays.^{48, 49} Vasta et al. examined the kinase profiling of crizotinib and dasatinib in live cells using NanoBRET based target engagement assay and found that the engagement of many kinases (71% for crizotinib and 50% for dasatinib) suggested by biochemical measurements were not confirmed in their live cell assay, which was likely due to the high concentration of local cellular ATP.⁴⁹ In summary, in vitro approaches (e.g., radioligand binding assay) give insight about the intrinsic properties of the ligands; however, it is more important to evaluate them with cellular approaches (e.g., NanoBRET^™ based target engagement assay) to more accurately predict their behaviors in vivo.

Every assay has its limitations. First, the coexpression of receptor with HiBiT and tagging of peptides with fluorophores via a non-natural linker might impact binding between the two partners. While the small 11-amino acid HiBiT tag is expected to have a minimal impact on the native function of GPCRs, the impact of the 19 kDa NanoBiT luciferase complex (formed after complementation with LgBiT) on the function of GPCRs remains uncertain. For the peptide-based tracers used in our experiments, we did notice slight alterations of pharmacological properties from their parent unmodified peptides. Second, in cell-based NanoBRET^™ target engagement assays antagonist-derived tracers are generally preferred to minimize activation and subsequent internalization of receptors. However, there are no small molecules or peptide antagonists available for GALR1, and we found that M40, which was previously reported as a non-selective galanin receptor antagonist,¹⁷ behaved as a full agonist in G-protein and arrestin assays and a partial agonist in our internalization assay.¹⁸ All peptide tracers used in our assay behaved as agonists of the β-arrestin pathway. Third, the foundation of the NanoBRET^™ assay is bioluminescence resonance energy transfer. Thus, compounds that interfere with the process of bioluminescence energy transfer might compromise results. For example, some compounds can non-specifically quench or enhance the fluorescence emitted by the acceptor fluorophore and alter the NanoBRET signal ratio. They can do this by either directly inhibiting the NanoLuc luciferase activity⁵⁰ or bind to stabilize and enhance the NanoLuc luciferase activity in cells. Other compounds can also absorb the bioluminescent light with dissipation through non-radiative pathways which would subsequently reduce BRET efficiency and signal output. These potential interferences must be carefully taken into account when interpreting NanoBRET^™ assay data.

Conclusion

In summary, we have developed a HiBiT peptide-based NanoBRET ligand binding assay for GALR1 in live cells. The binding affinity and the binding kinetics of six novel tracers on GALR1 were characterized in living cells. These tracers may also be useful for studying galanin interactions with other targets such as GAL2R. The assay was validated by the displacement of the tracers with a panel of unmodified galanin-derived peptides. Our study reiterates that the HiBiT peptide based NanoBRET^™ ligand binding assay is a valuable tool for studying the cellular target engagement of ligands, and for driving structure activity relationship studies in medicinal chemistry campaigns.

Supplementary Material

Supplementary figures, tables, structures and characterization of peptides/tracers, and supplementary LC/MS and HRMS characterization methods.