Advertisement

Label-free LC-MS based assay to characterize small molecule compound binding to cells

Open AccessPublished:September 02, 2022DOI:https://doi.org/10.1016/j.slasd.2022.08.005

      Abstract

      Study of small molecule binding to live cells provides important information on the characterization of ligands pharmacologically. Here we developed and validated a label-free, liquid chromatography-mass spectrometry (LC-MS) based cell binding assay, using centrifugation to separate binders from non-binders. This assay was applied to various target classes, with particular emphasis on those for which protein-based binding assay can be difficult to achieve. In one example, to study a G protein coupled receptor (GPCR), we used one antagonist as probe and multiple other antagonists as competitor ligands. Binding of the probe was confirmed to be specific and saturable, reaching a fast equilibrium. Competition binding analysis by titration of five known ligands suggested a good correlation with their inhibition potency. In another example, this assay was applied to an ion channel target with its agonists, of which the determined binding affinity was consistent with functional assays. This versatile method allows quantitative characterization of ligand binding to cell surface expressed targets in a physiologically relevant environment.

      Keywords

      Introduction

      Confirmation of compound binding to therapeutic targets plays an important role in drug discovery research [
      • Zhang R.
      • et al.
      Moderate to high throughput in vitro binding kinetics for drug discovery.
      ]. Traditionally, the binding is determined and quantified by assays using radioactive [
      • Maguire J.J.
      • Kuc R.E.
      • Davenport A.P.
      Radioligand Binding Assays and Their Analysis.
      ,
      • Bylund D.B.
      • Toews M.L.
      Radioligand Binding Methods for Membrane Preparations and Intact Cells.
      ] or fluorescent [
      • Emami-Nemini A.
      • et al.
      Time-resolved fluorescence ligand binding for G protein-coupled receptors.
      ,
      • Breen C.J.
      • Raverdeau M.
      • Voorheis H.P.
      Development of a quantitative fluorescence-based ligand-binding assay.
      ] label. However, labeling of compound can be laborious with some safety concerns (especially for radio-labeling), and some labels may affect compound binding to protein targets. All these factors compromise the wide use of labeled compounds in binding assays.
      Many important therapeutic targets belong to membrane associated proteins. These proteins are often challenging to produce or not in a relevant form when isolated [
      • Errey J.C.
      • Fiez-Vandal C.
      Production of membrane proteins in industry: The example of GPCRs.
      ,
      • Lin S.H.
      • Guidotti G.
      Purification of membrane proteins.
      ]. Even in membrane-associated preparations, the protein could be biased towards a state that lacks relevance. For example, many GPCRs require G proteins to maintain an active state [
      • Weis W.I.
      • Kobilka B.K.
      The Molecular Basis of G Protein-Coupled Receptor Activation.
      ]. On the other hand, the use of cells for binding assays maintains the protein in a more physiologically relevant environment and eliminates the need for membrane or protein preparation. Various technologies have been adapted to study ligand binding and/or protein-protein interactions on the surface of live cells, including the traditional radioligand binding assay [
      • Zeilinger M.
      • et al.
      New approaches for the reliable in vitro assessment of binding affinity based on high-resolution real-time data acquisition of radioligand-receptor binding kinetics.
      ,
      • Lobell R.B.
      • et al.
      A cell-based radioligand binding assay for farnesyl: protein transferase inhibitors.
      ], flow cytometry (FACS sorting) [
      • Jones M.L.
      • et al.
      Targeting membrane proteins for antibody discovery using phage display.
      ,
      • Uchański T.
      • et al.
      An improved yeast surface display platform for the screening of nanobody immune libraries.
      ,
      • Burgess T.L.
      • et al.
      A homogeneous SIRPα-CD47 cell-based, ligand-binding assay: Utility for small molecule drug development in immuno-oncology.
      ], and time-resolved FRET (TR-FRET) [
      • Stockmann H.
      • et al.
      Cell-Surface Receptor-Ligand Interaction Analysis with Homogeneous Time-Resolved FRET and Metabolic Glycan Engineering: Application to Transmembrane and GPI-Anchored Receptors.
      ]. Although being quantitative for some of them, all these approaches for the detection of specific binding require labelling or staining of the cells, targets, or ligands.
      There have been attempts in studying ligand binding to live cells. Both cellular thermal shift assays (CETSA) and surface plasma resonance (SPR) have been developed to study cellular target engagement [
      • Henderson M.J.
      • et al.
      High-Throughput Cellular Thermal Shift Assays in Research and Drug Discovery.
      ,
      • Mamer S.B.
      • et al.
      The Convergence of Cell-Based Surface Plasmon Resonance and Biomaterials: The Future of Quantifying Bio-molecular Interactions-A Review.
      ]. However, experimental setup for both assays is sophisticated. Moreover, CETSA is an indirect measurement of binding and may results in false positive results, so additional control experiments may be needed [
      • Savitski M.M.
      • et al.
      Tracking cancer drugs in living cells by thermal profiling of the proteome.
      ]; while cell-based SPR requires substantial optimization for immobilization. Another new assay to study target engagement is nanoBRET, but it requires co-expression of a fusion protein of target and luciferase and a labelled tracer [
      • Stoddart L.A.
      • Kilpatrick L.E.
      • Hill S.J.
      NanoBRET Approaches to Study Ligand Binding to GPCRs and RTKs.
      ,
      • Robers M.B.
      • et al.
      Quantitative, Real-Time Measurements of Intracellular Target Engagement Using Energy Transfer.
      ].
      Mass spectrometry-based binding assays have been proven capable of quantifying small molecule binding to protein targets without a requirement for compound labeling [
      • Hofstadler S.A.
      • Sannes-Lowery K.A.
      Applications of ESI-MS in drug discovery: interrogation of noncovalent complexes.
      ,
      • Muchiri R.N.
      • van Breemen R.B.
      Drug discovery from natural products using affinity selection-mass spectrometry.
      ,
      • Motoyaji T.
      Revolution of Small Molecule Drug Discovery by Affinity Selection-Mass Spectrometry Technology.
      ,
      • Bennett J.L.
      • Nguyen G.T.H.
      • Donald W.A.
      Protein-Small Molecule Interactions in Native Mass Spectrometry.
      ]. In addition, mass spectrometry is a straightforward and accurate method for quantifying ligand concentration. However, the reported applications have been limited to purified proteins or membrane-associated preparations [
      • Ackermann T.M.
      • et al.
      MS binding assays for GlyT1 based on Org24598 as nonlabelled reporter ligand.
      ,
      • Massink A.
      • et al.
      Mass spectrometry-based ligand binding assays on adenosine A1 and A2A receptors.
      ,
      • Sichler S.
      • et al.
      Development of MS Binding Assays targeting the binding site of MB327 at the nicotinic acetylcholine receptor.
      ,
      • Chen X.
      • et al.
      Label-Free, LC-MS-Based Assays to Quantitate Small-Molecule Antagonist Binding to the Mammalian BLT1 Receptor.
      ]. Therefore, there is still a need for methods that can quantitate ligand binding to their targets in a more physiologically relevant environment.
      Due to these challenges, development of a binding assay that uses label-free ligands and live cells would offer great advantages. Here we studied two different classes of targets, one GPCR (OX2R) and one ion channel receptor (TRPML1), and successfully developed and validated LC-MS based cell binding assays. For each target, characterization of probe binding to cells generated pharmacological parameters comparable to previous functional/binding assays. Our assay does not require labelling of any ligands or cells/targets, and it allows an efficient while physiologically relevant study of the target. This assay is applicable to various target classes.

      Results

      Method development of the LC-MS based cell binding assay

      To achieve a sensitive and accurate detection of bound ligand in a complex system of live cells, a highly sensitive triple quadrupole mass spectrometer was employed, and we applied targeted MS in selected reaction monitoring mode to ensure best accuracy. The LC effluent was directly coupled to the mass spectrometer. An internal standard was included in each sample for normalization purpose. For OX2R, compound MK-1064 [
      • Roecker A.J.
      • et al.
      Discovery of 5′'-chloro-N-[(5,6-dimethoxypyridin-2-yl)methyl]-2,2′:5′,3′'-terpyridine-3′-carboxamide (MK-1064): a selective orexin 2 receptor antagonist (2-SORA) for the treatment of insomnia.
      ] was used as internal standard (Fig. 1A). For TRPML1, a compound of similar structure to the probe was used.
      Fig 1
      Fig. 1(A) Structures of DORA-1 and MK-1064 (internal standard for DORA-1); (B) LC-MS standard curve for DORA-1 (N=3), ranges from 0.488nM-500nM (left), with a closer look of the curve from 0.488nM-31.25nM (right); (C) LC-MS standard curve for Compound A (N=3), ranges from 0.488nM-500nM (left), with a closer look of the curve from 0.488nM-31.25nM (right).
      Separation of the bound probe ligand from free ligand was achieved by quick centrifugation of cells. To prevent extensive dissociation of binders, cold buffers were used and samples were kept on ice whenever possible throughout the experiment. Final elution was mixed 1:1 with internal standard and samples were subjected to LC-MS analysis.

      Assay development and validation with known ligands and targets

      For assay development and validation we first picked a GPCR, Orexin receptor type 2 (OX2R), as the target. There are two orexin receptors OX1R and OX2R which are highly expressed in the human neural system. Binding of orexin neuropeptides to these two receptors promotes wakefulness in humans [
      • Sakurai T.
      The neural circuit of orexin (hypocretin): maintaining sleep and wakefulness.
      ,
      • Hong C.
      • et al.
      Structures of active-state orexin receptor 2 rationalize peptide and small-molecule agonist recognition and receptor activation.
      ]. A potent (IC50 4nM in FLIPR assay) dual orexin receptor antagonist (DORA-1) [
      • Bergman J.M.
      • et al.
      Proline bis-amides as potent dual orexin receptor antagonists.
      ] was used as the probe for OX2R (Fig. 1A).
      To determine the lower limit of detection (LLOD) of DORA-1 by MS, a standard curve ranging from 0.5-500nM was generated (Fig. 1B). Based on the standard curve, there was a linear relationship between the area response ratio and the DORA-1 concentration across the whole titration range. Therefore, we conclude that DORA-1 has a detection limit of less than 1nM (signal-to-noise ratio >100) and can serve as a probe ligand for the assay. Being able to detect the probe at low concentrations ensures that we can accurately define probe bound below Kd, especially when we are limited by target expression level and the number of cells. With detection limit of ∼1 nM, 2 million cells and 1 ml incubation volume, minimum number of target protein per cell is 3 × 105. Expression of some endogenous proteins may fall below this limit.
      Then we moved on to confirm the specific binding of DORA-1 to OX2R cells. DORA-1 was incubated with OX2R cells to measure total binding. Two blocking compounds that bind to the same site of OX2R as DORA-1, MK-6096 [
      • Coleman P.J.
      • et al.
      Discovery of [(2R,5R)-5-{[(5-fluoropyridin-2-yl)oxy]methyl}-2-methylpiperidin-1-yl][5-methyl-2-(pyrimidin-2-yl)phenyl]methanone (MK-6096): a dual orexin receptor antagonist with potent sleep-promoting properties.
      ] and MK-3697 [
      • Roecker A.J.
      • et al.
      Discovery of MK-3697: a selective orexin 2 receptor antagonist (2-SORA) for the treatment of insomnia.
      ] were included (separately) in excess to measure non-specific binding (NSB). Both MK-6096 and MK-3697 are antagonists for OX2R. The difference between total binding and NSB was examined to confirm specific binding and define the specific binding window. The same was also performed with parental CHO-K1 cells, to confirm that DORA-1 binding is specific to OX2R. As shown in Fig. 2A, there was no specific binding of DORA-1 to parental cells, only a low level of NSB was observed. Both MK-6096 and MK-3697 could effectively block specific binding of DORA-1 to OX2R, when present in excess (Fig. 2B). MK-6096 was used as the blocker in all following experiments for OX2R.
      Fig 2
      Fig. 2Detection of probe binding to parental and target cells, N=2. For OX2R, effects of cell wash are compared between 3-wash (open) and no-wash (dotted) method with parental CHO-K1 cells (A) and OX2R cells (B); Specific binding of DORA-1 to OX2R cells between using 2 million cells (open) and 5 million cells (striped) were examined and compared (C); Binding of Compound A to parental HEK293 cells (patterned) and TRPML1 (open) were examined (D).
      Various assay conditions were then tested, i.e. the number of cell washes and the cell number for incubation. DORA-1 showed the strongest level of specific binding after 3 washes of cells, with a 297-fold difference between total binding and NSB (in the presence of excess MK-6096), compared to the no-wash method (8.5-fold difference) (Fig. 2B). For parental CHO-K1 cells, signals of both total binding and NSB are comparable to the NSB in OX2R cells, suggesting little to no specific binding (Fig. 2A). In this experiment, 2 million cells were used for incubation. Increasing the cell number (2 million to 5 million) resulted in higher background and reduced the specific binding window. As shown in Fig. 2C, the difference between total binding of DORA-1 and NSB (in the presence of excess MK-6096) decreased when using 5 million cells (133-fold), compared to 2 million cells (310-fold), under the same condition. Based on these results, 3-wash method with 2 million cells was selected as the final assay condition.
      TRPML1, an ion channel receptor in the lysosome associated with exocytosis and autophagy [
      • Di Paola S.
      • Scotto-Rosato A.
      • Medina D.L.
      TRPML1: The Ca((2+))retaker of the lysosome.
      ,
      • Santoni G.
      • et al.
      Pathophysiological Role of Transient Receptor Potential Mucolipin Channel 1 in Calcium-Mediated Stress-Induced Neurodegenerative Diseases.
      ], was picked as a second example. In this case we used an engineered mutant cell line overexpressing TRPML1 at the cell surface membrane [
      • Zhang X.
      • Li X.
      • Xu H.
      Phosphoinositide isoforms determine compartment-specific ion channel activity.
      ]. An agonist ligand Compound A (EC50 24nM in FLIPR assay) was used as probe. It was confirmed to have a good lower limit of detection (LLOD) of less than 1nM ((signal-to-noise ratio >50) (Fig. 1C). Agonist compound B (EC50 27nM in FLIPR assay) was used in excess to measure NSB. Specific binding of the probe to TRPML1 cells was confirmed with 2 million cells and the 3-wash method, with a 7.9-fold difference between total binding and NSB (Fig. 2D).

      Direct saturation binding of probe to target cells

      With the assay established, equilibrium saturation binding of DORA-1 was performed to determine its Kd for OX2R cells. It was confirmed that binding of the probe reached equilibrium within 10min. Thus, cells were incubated for 30min with increasing concentrations of the probe (0.5-50nM) in the presence of an excess amount of MK-6096 (for NSB) and absence of MK-6096 (for total binding). In Fig. 3A, DORA-1 showed negligible level of NSB and thus specific binding was close to total binding, with a Kd of 12.6nM. This Kd was on par with the IC50 of DORA-1 (4nM) in FLIPR assay.
      Fig 3
      Fig. 3Equilibrium saturation binding of the probe was performed to determine its Kd and Bmax for OX2R cells (A, N=2) and TRPML1 cells (B, N=1). DORA-1 Kd was measured as 12.6nM (IC50 is 4nM from FLIPR assay) and Bmax as 6.8nM/million cells. Compound A Kd was measured as 22nM (EC50 is 24nM from FLIPR assay) and Bmax as 1.1nM/million cells. curves were fitted of the mean.
      Based on equilibrium saturation binding of Compound A to TRPML1 cells (Fig. 3B), a Kd of 22nM was calculated for Compound A. This correlates well with functional assay results (EC50 24nM in FLIPR assay). Similar to OX2R, the experiment was performed by incubating TRPML1 cells with increasing concentrations of Compound A (5-160nM). NSB was measured in the presence of excess amount of Compound B and specific binding was calculated as total binding subtracting NSB.

      Dissociation binding studies of the probe

      The kinetics of ligand binding provides important pharmacological information. To study the binding kinetics DORA-1 to OX2R cells, the dissociation rate (koff) of DORA-1 was determined. OX2R cells were first saturated with the probe, before an excess amount of competitor MK-6096 was added to initiate dissociation. Dissociation was stopped at different time points and the level of the probe remaining bound to cells was quantified by LC-MS (Fig. 4). The dissociation rate koff was calculated to be 0.18 min−1 for DORA-1, corresponding to a half-life (t1/2) of about 3.8min. This data suggests that binding of DORA-1 to OX2R cells is reversible and reaches equilibrium rapidly. With detection limit of ∼1 nM and total 6-minute process time, the theoretical detection limit of dissociation constant is calculated to be 0.018min−1 based on dissociation rate equation.
      Fig 4
      Fig. 4Dissociation of DORA-1 from OX2R cells in time course, each data point represents mean ± error (N=2). Dissociation rate koff was calculated to be 0.18 min−1, corresponding to a half-life (t1/2) of about 3.8min.

      Competition binding of ligands to target cells with the probe

      A common method for evaluation of the affinity of a ligand is to perform a competition binding assay with a known probe. Here for OX2R, we selected five antagonists with different potencies (MK-6096, MK-3697, Compound C, D, E) [
      • Roecker A.J.
      • et al.
      Discovery of MK-3697: a selective orexin 2 receptor antagonist (2-SORA) for the treatment of insomnia.
      ,
      • Whitman D.B.
      • et al.
      Discovery of a potent, CNS-penetrant orexin receptor antagonist based on an n,n-disubstituted-1,4-diazepane scaffold that promotes sleep in rats.
      ] to compete with the probe (Fig. 5A). In Table 1 their Ki and IC50 values that had been measured previously are listed. To evaluate their potencies with our assay, OX2R cells were pre-incubated with increasing concentrations of the competitor ligand and then incubated with subsaturating amounts of DORA-1. The level of bound DORA-1 was quantified by LC-MS and plotted against each competitor concentration (Fig. 5). Based on the results, all five competitors were able to compete off the probe, in a concentration-dependent manner. IC50 was derived from each dissociation curve (Fig. 5B), and Ki of each competitor was determined using GraphPad Prism 8. The weakest compound affinity that we could detect is 4.6µM (compound E). The results are summarized in the last column of Table 1.
      Fig 5
      Fig. 5(A) Structures of MK-6096, MK-3697, and Compound C, D, E for OX2R; (B) Determination of Ki for MK-6096, MK-3697 and Compound C, D, E, each data point represents mean ± error (N=2), with curve fitting of the mean; (C) Determination of Ki for Compound B of TRPML1 (N=2).
      Table 1Determined Ki from previous competition binding assay and from our approach.
      Competitor ligandsDetermined Ki from radioligand binding assay
      Data represent mean (N>=2)
      IC50 from FLIPR assay
      Data represent mean (N>=2)
      Determined Ki from cell binding assay*
      MK-60960.3nM11nM8.8nM
      MK-36971.1nM16nM46nM
      Compound C11nM120nM324nM
      Compound D38nM215nM1036nM
      Compound E193nM1450nM4630nM
      low asterisk Data represent mean (N>=2)
      The competitor ligands we chose have a wide range of potency, with IC50 in FLIPR assay ranging from 11nM to more than 1uM. Our cell binding assay data showed that we were able to evaluate the potency of all of them with relatively good accuracy. Most importantly, the ranking order of competition binding potency correlated with that from radioligand binding assay and FLIPR assay. These results validate this newly developed cell binding assay.
      For TRPML1, competition binding of Compound B with Compound A was performed. Cells were preincubated with increasing concentrations of Compound B and then incubated with subsaturating amounts of Compound A. Based on the dissociation curve (Fig. 5C), Ki of Compound B was calculated to be 27nM, suggesting a good correlation with previously determined potency (EC50 27nM in FLIPR assay). In addition, the fact that Compound B's potency measured here is very close to Compound A's potency (based on direct saturation binding), is consistent with the conclusion that the two compounds have comparable potencies drawn from functional FLIPR assay. This adds confidence to our data. Together the data on both targets indicate that the LC-MS based cell binding assay is applicable to targets of different classes.

      Discussion

      In this paper we described the development and validation of an LC-MS based cell binding assay, with two targets of different classes shown as examples. With this assay we were able to perform quantitative analysis on ligand binding to cell surface expressed targets. This assay eliminates the need for labeling of any ligands or protein targets, as well as the need for protein production and activity confirmation. In addition, our results show that this assay is applicable to a wide range of targets of different classes. For both targets studied here, the pharmacological parameters calculated from our assay are generally in good agreement with those from previous functional assays, indicating that our assay could be a method for quantitative characterization of target-ligand binding. There is a left shift in potency for radioligand binding assay, compared to both our assay and the functional assay (FLIPR assay), which may result from the use of membranes versus live cells. There is also a slight difference in compound potency between our assay that measures compound binding and the FLIPR assay that measures the effect on cell function. Nevertheless, for most of compounds that we tested this difference is less than 3-fold. Although both targets are on cell surface, we expect this approach can be applied to intracellular targets; however, cell permeability of tested compounds will be a contributing factor to signal intensity in addition to binding affinity.
      This assay is a good addition to our current assay toolbox of small molecule hit confirmation and characterization. Together with other functional and binding assays, they provide a more comprehensive characterization of hits in vitro. With the emergence of various affinity-based selection technologies for small molecules such as DNA Encoded Library (DEL) [
      • Satz A.L.
      • et al.
      DNA-encoded chemical libraries.
      ], as well as various kinds of display technologies for screening of peptides and antibodies (mRNA display etc.) [
      • Huang Y.
      • Wiedmann M.M.
      • Suga H.
      RNA Display Methods for the Discovery of Bioactive Macrocycles.
      ], there is a strong need for a quick and robust reagent validation and hit characterization process. Our assay therefore serves as a good strategy for both reagent QC and hit confirmation, including small molecules and beyond (peptides, antibodies, etc.).
      For cells we used stable cell line for OX2R and inducible cell line for TRPML1. Although transient cell line is likely to provide higher expression level of the target protein, strict experimental setup is required to avoid varying level of expression due to plasmid lost, which is a less-controlled process. In case stable overexpression of target protein is a concern, an inducible cell line may be an alternative option.
      In the competition binding assay mode, we were able to rank order unknown compounds in the presence of a potent known probe, with unknown compound potency detection limit ranging from single-digit nM to single-digit uM. Such a wide range makes our approach a powerful tool for hit confirmation. However, one limitation is the prerequisite of a potent probe. Depending on the target expression level and sensitivity of the mass spectrometer, and based on our experience, a double-digit nM probe would be needed for this approach. In the direct binding mode, a probe is not required, though MS detection method has to be established for each compound.
      This LC-MS based cell binding assay has a potential to be developed into a high-throughput format, although certain modifications are required. For example, multi-well plate and less reagent can be used to miniaturize the assay, and cell wash can be automated. However, these modifications will not be straightforward without technical challenges. With current approach of 2 million cells per sample, converting to 384-well format means 800 million cells for only one plate, which is unrealistic logistically. In addition, 1 ml used in wash steps is more than well volume in 384-well plate. Effective centrifugation of multi-well plates can also pose a technical challenge. All these factors means extensive work and major changes are needed to miniaturize into a high-throughput format.
      With all the future development, the assay will have even broader applications and be readily applied to routine hit confirmation. Furthermore, with the ability to screen large numbers of compounds, based on screening data of percentage of binders out of screened compounds, ligandability of a target can be assessed by the likelihood of finding small molecule binders to the target. In addition, probe compounds can be discovered and utilized to validate the target for therapeutic purpose.

      Materials and methods

      Materials

      DPBS (no calcium, no magnesium) was purchased from GIBCO/Thermo Fisher Scientific (Waltham, MA); DMSO was purchased from Sigma-Aldrich (St. Louis, MO); all cell culture media and reagents were purchased from GIBCO/Thermo Fisher Scientific (Waltham, MA). CHO-K1 (CCL-61) cell line was from ATCC (Manassas, VA). HEK293 Flp-In™ T-REx™ cell line was from Invitrogen/Thermo Fisher Scientific (Waltham, MA). All compounds tested were synthesized, purified, and authenticated by the Research Laboratories of Merck & Co., Inc., Rahway, NJ, USA.
      OX2R cells (CHO hOX2R NFAT Clone 7) were produced as described in [
      • Cox C.D.
      • et al.
      Discovery of the dual orexin receptor antagonist [(7R)-4-(5-chloro-1,3-benzoxazol-2-yl)-7-methyl-1,4-diazepan-1-yl][5-methyl-2-(2H-1,2,3-triazol-2-yl)phenyl]methanone (MK-4305) for the treatment of insomnia.
      ]. Cells were grown in Iscove's Modified Dulbecco's Medium (IMDM) containing 10% heat-inactivated fetal bovine serum (FBS), 500ug/mL G418, and 100U/mL Penicillin-Streptomycin (Pen Strep). Cells were cultured in T175 flasks at 37°C with 5% CO2, and passaged at 80-90% confluency using TrypLE Express Enzyme.
      The surface-expressing mutant of human TRPML1 (NM_020533.2) was constructed as described in [
      • Zhang X.
      • Li X.
      • Xu H.
      Phosphoinositide isoforms determine compartment-specific ion channel activity.
      ]. cDNA sequence was cloned into pcDNA™5⁄FRT⁄TO vector, and the recombinant plasmid was transfected into HEK293 Flp-In™ T-REx™ host cell line. Stable clones were selected with Hygromycin. TRPML1 cells were grown in Dulbecco's modified Eagle medium (DMEM) containing 10% heat-inactivated Tetracycline-free FBS, 1X MEM Non-Essential Amino Acids (NEAA) Solution, 2mM L-Glutamine, 100U/mL Penicillin-Streptomycin (Pen Strep), 15ug/mL Blasticidin S HCl, 200ug/mL Hygromycin. To induce TRPML1 expression, Tetracycline was added at a final concentration of 2ug/mL to medium and cells were induced overnight. Cells were cultured in T175 flask at 37°C with 5% CO2, and passaged at 80-90% confluency using TrypLE Express Enzyme.

      Methods

      Determination of the lower detection limit of probe

      Serial dilutions of the probe (DORA-1 or Compound A) were prepared in DPBS from 0.976nM to 1000nM. They were mixed 1:1 with the corresponding internal standard (200nM in DPBS), so that the final titration range is 0.488nM to 500nM. Samples were analyzed by LC-MS.

      Cell binding assay

      To study the binding of probe to target cells, 2 million cells were collected and incubated with the probe in DPBS in a total volume of 1mL. For measuring NSB, MK-6096 (or MK-3697) was added at a final conc. of 10uM for OX2R, and Compound B was added at a final conc. of 20uM for TRPML1. To be consistent, the same volume of DMSO was added to samples for measuring total binding. To ensure effective blocking, the incubation was kept at 37°C for 10min before any probe was added (DORA-1 final conc. 10nM, Compound A final conc. 25nM). All incubations were then maintained at 37°C for 30min with slow rotation. Cells were collected by centrifugation at 352 g-force for 3min and washed with 1mL ice-cold DPBS 3 times. All samples were kept on ice whenever possible during the process. After final wash cells were resuspended in 100uL DPBS and heated at 95°C for 10min, to release bound ligands. The solution was centrifuged at 8765 g-force for 3min and supernatant was taken. This final elution was mixed 1:1 with internal standard and analyzed by LC-MS. The same experiments were performed with parental cells as well. All experiments were done in duplicate.
      For no-wash method, OX2R cells were collected after incubation and directly resuspended in 100uL DPBS for heating, while all other steps remain the same. For cell number comparison, the experiments were performed with 5 million OX2R cells with all the steps being the same as those for 2 million cells, except that cells were incubated with 25nM DORA-1 and the blocker final concentration was increased to 25uM when needed.

      Study of equilibrium saturation binding

      Saturation binding of DORA-1 was performed with 2 million OX2R cells and increasing concentration of DORA-1 (0.5, 1, 2, 4, 8, 16, 32, 50nM). To determine NSB, cells were pre-incubated with 10uM MK-6096 for 10min at 37°C. Total binding was defined as binding in the absence of MK-6096. Binding was initiated by adding the indicated amount of DORA-1 to each sample. Same as above, cells were washed 3 times and heated to get the final elution, which was mixed with MK-1064 and subject to LC-MS analysis. Specific binding was calculated as total binding minus NSB. All data points were plotted in GraphPad Prism 8 (GraphPad Software, San Diego, CA). Kd was calculated by fitting the data with one-site specific binding model.
      The same experiment was performed for Compound A of TRPML1, with 2 million TRPML1 cells and increasing conc. of Compound A (0, 5, 20, 40, 80, 160nM). 20uM Compound B was used as blocker to measure NSB.

      Study of dissociation binding

      To study DORA-1 dissociation kinetics, 10nM DORA-1 was first incubated with 2 million OX2R cells for 30min at 37°C, for binding to reach equilibrium. Dissociation was initiated by adding 10uM MK-6096 and incubated for different time periods. To ensure best accuracy, the experiments were performed in a reversed-time manner. The level of bound DORA-1 before adding MK-6096 (t=0) was defined as total binding. The level of DORA-1 remaining bound at each time point was divided by the total binding level to generate the % Bound. Data were plotted and koff and half-life were calculated by fitting the data to one phase exponential decay model for dissociation studies.

      Study of competition binding

      For competition binding, the competitor ligand at different concentration was pre-incubated with 2 million target cells for 10min at 37°C, before the probe was added to each sample to initiate competition (10nM DORA-1, or 25nM Compound A). The incubation was stopped after 30min and cells were treated the same as before. Bound probe was quantified by LC-MS and data points were plotted in GraphPad Prism 8 (GraphPad Software, San Diego, CA). IC50 and Ki was generated by fitting the data with one-site fit IC50 and one-site fit Ki model.

      LC-MS analysis

      A Thermo TSQ Vantage triple quadrupole mass spectrometer with electrospray ionization source was employed for MS analysis. A Thermo Accucore C18 HPLC column (30 × 2.1 mm, 2.6um particle size) and water/acetonitrile binary gradient were employed for LC separation. The separation conditions were the same as [
      • Chen X.
      • et al.
      Label-Free, LC-MS-Based Assays to Quantitate Small-Molecule Antagonist Binding to the Mammalian BLT1 Receptor.
      ], with mobile phase A being water/0.1% formic acid, mobile phase B being acetonitrile/0.1% formic acid. Acquisition starts with 15sec of column equilibration, followed by gradient ramping from 15% B to 95% B in 4min and holding for 1min; the gradient then stepped down to 15% B (15sec) and was held there for 0.5min. The total acquisition time is 6min. For both DORA-1 (OX2R) and Compound A (TRPML1), positive ion mode was used. For DORA-1, the following MS operating parameters were used: spray voltage 3000 V, sheath gas pressure 50 arbitrary units, auxiliary gas pressure 55 arbitrary units, vaporizer temperature 350°C, capillary temperature 270°C, and collision pressure 1.5 mTorr. The S-lens and collision energy for DORA-1 were set as 109 and 30V, with a transition from 471.0 (m/z) to 204.9 (m/z) monitored. As for internal standard MK-1064, the S-lens and collision energy were set as 85 and 16V, with a transition from 462.0 (m/z) to 294.0 (m/z) monitored. For Compound A, the following MS operating parameters were used: spray voltage 3500 V, sheath gas pressure 10 arbitrary units, auxiliary gas pressure 55 arbitrary units, vaporizer temperature 350°C, capillary temperature 270°C, and collision pressure 1.5 mTorr. The S-lens and collision energy for Compound A were set as 120 and 32V, with a transition from 452.1 (m/z) to 238.0 (m/z) monitored. While the internal standard for Compound A has a S-lens of 114 and collision energy of 31V, and a transition from 430.0 (m/z) to 226.0 (m/z). Data were processed using Thermo LCquan.

      FLIPR assay

      For OX2R, FLIPR assays were performed following the same protocol as in [
      • Cox C.D.
      • et al.
      Discovery of the dual orexin receptor antagonist [(7R)-4-(5-chloro-1,3-benzoxazol-2-yl)-7-methyl-1,4-diazepan-1-yl][5-methyl-2-(2H-1,2,3-triazol-2-yl)phenyl]methanone (MK-4305) for the treatment of insomnia.
      ].
      For TRPML1, FLIPR calcium 6 Assay Kit (Molecular Devices, San Jose, CA) was used for EC50 measurement of TRPML1 agonists. TRPML1 cells were seeded at 4000 cells/well in 4uL assay media (DMEM containing 1% FBS, 1X NEAA, 100U/ml Pen Strep, 2ug/mL Tetracycline) to 1536 PDL coated black clear plates (Aurora, Whitefish MT) and incubated overnight. Agonist compound was prepared in assay buffer from 30uM at ½ log titration for a total of 7-point measurement. The next day, 3uL calcium 6 loading dye resuspended in FLIPR assay buffer (1x HBSS, 20 mM HEPES, 2.5mM probenecid, 0.1% BSA) was loaded to each well of the plate and incubated for 2 hours. After that, 3uL of compound was added to each well the cell plate using FLIPR Tetra (Molecular Devices, San Jose, CA) and signals were recorded. Stimulation in % is calculated as the ratio of signal after stimulation versus signal before stimulation. Data was processed and analyzed by Spotfit in Spotfire to generate EC50, using 4 parameter logistic fit based on the Levenberg-Marquardt algorithm.

      Radioligand binding assay

      Ki of various OX2R antagonists were determined by radioligand binding assay as described in [
      • Cox C.D.
      • et al.
      Discovery of the dual orexin receptor antagonist [(7R)-4-(5-chloro-1,3-benzoxazol-2-yl)-7-methyl-1,4-diazepan-1-yl][5-methyl-2-(2H-1,2,3-triazol-2-yl)phenyl]methanone (MK-4305) for the treatment of insomnia.
      ]. A tritiated ligand structurally similar to DORA-1 was used in the experiment.

      Declaration of Conflicting Interests

      The authors have declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: All authors are, or were, employed by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and their research and authorship of this article was completed within the scope of their employment.
      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgements

      The authors would like to thank Xi Ai, Ryan Loy, Robert Ramos, Qifeng Yang, Naman Patel, Andrea Peier for their support on cell line generation and cell culture, Michael Rudd, Terrence Mcdonald, Charles Meacham Harrell, Scott Mosser, and Kasper Hollenstein for their input on OX2R and Mark Fraley, Brain Magliaro, Rob Drolet for their support on TRPML1.

      Funding

      The authors received no financial support for the research, authorship, and/or publication of this article.

      References

        • Zhang R.
        • et al.
        Moderate to high throughput in vitro binding kinetics for drug discovery.
        Front Biosci. 2016; 8 (Schol Ed): 278-297
        • Maguire J.J.
        • Kuc R.E.
        • Davenport A.P.
        Radioligand Binding Assays and Their Analysis.
        in: Davenport A.P Receptor Binding Techniques. Humana Press, Totowa, NJ2012: 31-77 (Editor)
        • Bylund D.B.
        • Toews M.L.
        Radioligand Binding Methods for Membrane Preparations and Intact Cells.
        in: Willars G.B. Challiss R.A.J. Receptor Signal Transduction Protocols. Third Edition. Humana Press, Totowa, NJ2011: 135-164 (Editors)
        • Emami-Nemini A.
        • et al.
        Time-resolved fluorescence ligand binding for G protein-coupled receptors.
        Nat Protoc. 2013; 8: 1307-1320
        • Breen C.J.
        • Raverdeau M.
        • Voorheis H.P.
        Development of a quantitative fluorescence-based ligand-binding assay.
        Sci Rep. 2016; 6: 25769
        • Errey J.C.
        • Fiez-Vandal C.
        Production of membrane proteins in industry: The example of GPCRs.
        Protein Expr Purif. 2020; 169105569
        • Lin S.H.
        • Guidotti G.
        Purification of membrane proteins.
        Methods Enzymol. 2009; 463: 619-629
        • Weis W.I.
        • Kobilka B.K.
        The Molecular Basis of G Protein-Coupled Receptor Activation.
        Annu Rev Biochem. 2018; 87: 897-919
        • Zeilinger M.
        • et al.
        New approaches for the reliable in vitro assessment of binding affinity based on high-resolution real-time data acquisition of radioligand-receptor binding kinetics.
        EJNMMI Res. 2017; 7: 22
        • Lobell R.B.
        • et al.
        A cell-based radioligand binding assay for farnesyl: protein transferase inhibitors.
        J Biomol Screen. 2003; 8: 430-438
        • Jones M.L.
        • et al.
        Targeting membrane proteins for antibody discovery using phage display.
        Sci Rep. 2016; 6: 26240
        • Uchański T.
        • et al.
        An improved yeast surface display platform for the screening of nanobody immune libraries.
        Sci Rep. 2019; 9: 382
        • Burgess T.L.
        • et al.
        A homogeneous SIRPα-CD47 cell-based, ligand-binding assay: Utility for small molecule drug development in immuno-oncology.
        PLoS One. 2020; 15e0226661
        • Stockmann H.
        • et al.
        Cell-Surface Receptor-Ligand Interaction Analysis with Homogeneous Time-Resolved FRET and Metabolic Glycan Engineering: Application to Transmembrane and GPI-Anchored Receptors.
        J Am Chem Soc. 2017; 139: 16822-16829
        • Henderson M.J.
        • et al.
        High-Throughput Cellular Thermal Shift Assays in Research and Drug Discovery.
        SLAS Discov. 2020; 25: 137-147
        • Mamer S.B.
        • et al.
        The Convergence of Cell-Based Surface Plasmon Resonance and Biomaterials: The Future of Quantifying Bio-molecular Interactions-A Review.
        Ann Biomed Eng. 2020; 48: 2078-2089
        • Savitski M.M.
        • et al.
        Tracking cancer drugs in living cells by thermal profiling of the proteome.
        Science. 2014; 3461255784
        • Stoddart L.A.
        • Kilpatrick L.E.
        • Hill S.J.
        NanoBRET Approaches to Study Ligand Binding to GPCRs and RTKs.
        Trends Pharmacol Sci. 2018; 39: 136-147
        • Robers M.B.
        • et al.
        Quantitative, Real-Time Measurements of Intracellular Target Engagement Using Energy Transfer.
        in: Ziegler S. Waldmann H. Systems Chemical Biology: Methods and Protocols. Springer, New York: New York, NY2019: 45-71 (Editors)
        • Hofstadler S.A.
        • Sannes-Lowery K.A.
        Applications of ESI-MS in drug discovery: interrogation of noncovalent complexes.
        Nat Rev Drug Discov. 2006; 5: 585-595
        • Muchiri R.N.
        • van Breemen R.B.
        Drug discovery from natural products using affinity selection-mass spectrometry.
        Drug Discov Today Technol. 2021; 40: 59-63
        • Motoyaji T.
        Revolution of Small Molecule Drug Discovery by Affinity Selection-Mass Spectrometry Technology.
        Chem Pharmaceut Bull. 2020; 68: 191-193
        • Bennett J.L.
        • Nguyen G.T.H.
        • Donald W.A.
        Protein-Small Molecule Interactions in Native Mass Spectrometry.
        Chem Rev. 2022; 122: 7327-7385
        • Ackermann T.M.
        • et al.
        MS binding assays for GlyT1 based on Org24598 as nonlabelled reporter ligand.
        Neuropharmacology,. 2019; 161107561
        • Massink A.
        • et al.
        Mass spectrometry-based ligand binding assays on adenosine A1 and A2A receptors.
        Purinergic Signal. 2015; 11: 581-594
        • Sichler S.
        • et al.
        Development of MS Binding Assays targeting the binding site of MB327 at the nicotinic acetylcholine receptor.
        Toxicol Lett. 2018; 293: 172-183
        • Chen X.
        • et al.
        Label-Free, LC-MS-Based Assays to Quantitate Small-Molecule Antagonist Binding to the Mammalian BLT1 Receptor.
        SLAS Discov. 2017; 22: 1131-1141
        • Roecker A.J.
        • et al.
        Discovery of 5′'-chloro-N-[(5,6-dimethoxypyridin-2-yl)methyl]-2,2′:5′,3′'-terpyridine-3′-carboxamide (MK-1064): a selective orexin 2 receptor antagonist (2-SORA) for the treatment of insomnia.
        ChemMedChem. 2014; 9 (-22): 311
        • Sakurai T.
        The neural circuit of orexin (hypocretin): maintaining sleep and wakefulness.
        Nat Rev Neurosci. 2007; 8: 171-181
        • Hong C.
        • et al.
        Structures of active-state orexin receptor 2 rationalize peptide and small-molecule agonist recognition and receptor activation.
        Nat Commun. 2021; 12: 815
        • Bergman J.M.
        • et al.
        Proline bis-amides as potent dual orexin receptor antagonists.
        Bioorg Med Chem Lett. 2008; 18: 1425-1430
        • Coleman P.J.
        • et al.
        Discovery of [(2R,5R)-5-{[(5-fluoropyridin-2-yl)oxy]methyl}-2-methylpiperidin-1-yl][5-methyl-2-(pyrimidin-2-yl)phenyl]methanone (MK-6096): a dual orexin receptor antagonist with potent sleep-promoting properties.
        ChemMedChem. 2012; 7 (337): 415-424
        • Roecker A.J.
        • et al.
        Discovery of MK-3697: a selective orexin 2 receptor antagonist (2-SORA) for the treatment of insomnia.
        Bioorg Med Chem Lett. 2014; 24: 4884-4890
        • Di Paola S.
        • Scotto-Rosato A.
        • Medina D.L.
        TRPML1: The Ca((2+))retaker of the lysosome.
        Cell Calcium. 2018; 69: 112-121
        • Santoni G.
        • et al.
        Pathophysiological Role of Transient Receptor Potential Mucolipin Channel 1 in Calcium-Mediated Stress-Induced Neurodegenerative Diseases.
        Front Physiol. 2020; 11: 251
        • Zhang X.
        • Li X.
        • Xu H.
        Phosphoinositide isoforms determine compartment-specific ion channel activity.
        Proc Natl Acad Sci U S A,. 2012; 109: 11384-11389
        • Whitman D.B.
        • et al.
        Discovery of a potent, CNS-penetrant orexin receptor antagonist based on an n,n-disubstituted-1,4-diazepane scaffold that promotes sleep in rats.
        ChemMedChem. 2009; 4: 1069-1074
        • Satz A.L.
        • et al.
        DNA-encoded chemical libraries.
        Nat Rev Methods Primers. 2022; 2: 3
        • Huang Y.
        • Wiedmann M.M.
        • Suga H.
        RNA Display Methods for the Discovery of Bioactive Macrocycles.
        Chem Rev. 2019; 119: 10360-10391
        • Cox C.D.
        • et al.
        Discovery of the dual orexin receptor antagonist [(7R)-4-(5-chloro-1,3-benzoxazol-2-yl)-7-methyl-1,4-diazepan-1-yl][5-methyl-2-(2H-1,2,3-triazol-2-yl)phenyl]methanone (MK-4305) for the treatment of insomnia.
        J Med Chem. 2010; 53: 5320-5332