Development of Scintillation Proximity Assay for GPR23

GPR23 (P2Y9) is a novel receptor that binds lysophosphatidic acid (LPA), although its physiological importance is not fully understood (Noguchi, Ishii, and Shimizu 2003). In order to develop a GPR23 binding assay, we prepared membranes from a Chinese hamster ovary (CHO) cell line that expressed human GPR23 under the control of a tetracycline-inducible expression system. Under normal conditions, these cells expressed GPR23 at a very low level but could be induced by tetracy-cline (or a derivative) to express the receptor at a high level.

Expression of GPR23 was induced by treating the cells overnight with doxycycline (1 pg/mL). The next day, the cells were harvested and centrifuged at 1900 g for 10 min at 4°C to produce a cell pellet, which was washed with ice-cold phosphate-buffered saline (PBS) and suspended in a homog-enization buffer containing 25 mM HEPES (pH 7.4), 10% sucrose, and EDTA-free Complete™ protease inhibitor (Roche Diagnostics). The cells were left on ice for 30 min before being homogenized using a glass homogenizer. The cell lysate was centrifuged at 1300 g for 10 min at 4°C. The supernatant was collected and then centrifuged at 142,000 g for 1 hr at 4°C to produce a crude membrane particulate fraction. This fraction was suspended in a storage buffer containing 25 mM HEPES

(pH 7.4), 10 mM MgCl2, and EDTA-free Complete™ protease inhibitor and was then homogenized. The protein content of the membrane preparation was assessed using the DC protein assay (BioRad Laboratories) and subsequently the membranes were divided into aliquots containing >1 mg/ mL of protein.

The GPR23 binding SPA was developed in a 384-well format with luminescence measured on a TopCount system. Reagents were added into white flat-bottomed polystyrene plates (Corning, #3652) in the following order: 10 pL unlabeled compound in assay buffer, 20 pL premixed beads and cell membranes, and 10 pL [3H]LPA (Figure 4.2A). The composition of the assay buffer was similar to that of the storage buffer, but also included 0.2% bovine serum albumin (BSA). Several different

GPR23 Binding SPA

384-well plate containing compounds (10 |L)

Add SPA beads and cell membrane mixture (20 |iL)

J Incubate at room temp for 0.5 hr

SPA Bead and Membrane Protein Ratio

^ Incubate at room temp Read plate on TopCount®

H]-LPA Saturation Binding

700

600

)

500

m

pc

400

st tn

u

300

o

C

200

100

I Total

co Ln

Membrane Proteins (|g/well)

H]-LPA Saturation Binding

Concentration of [3H]-LPA (nM)

500 D

250 125

LPA Competition Binding

62.5

250 125

LPA Competition Binding

Log [LPA] (M) Scatterplot of a DMSO Plate

Concentration of [3H]-LPA (nM)

DMSO Effects on [3H]-LPA Binding

Log [LPA] (M) Scatterplot of a DMSO Plate

DMSO Effects on [3H]-LPA Binding

0 0.01 0.05 0.1 0.5 1 2 Final DMSO Concentration (%)

100 200 300

Well Number

FIGURE 4.2 Development of SPA-based ligand binding for GPR23. (A) Protocol for GPR23 binding assay. (B) Titration of beads and GPR23 membrane protein to determine optimal mixture ratio. (C) Saturation binding of [3H]LPA to GPR23. (D) Competitive displacement of [3H]LPA by unlabeled LPA. (E) Determination of DMSO tolerance of GPR23 assay. (F) Scatter plot of signals obtained in each well of 384-well plate. All wells in columns 1 through 23 contained 1.25% DMSO and the 16 wells in column 24 contained DMSO and 10 |M LPA before addition of mixed beads and membrane.

0 0.01 0.05 0.1 0.5 1 2 Final DMSO Concentration (%)

100 200 300

Well Number

FIGURE 4.2 Development of SPA-based ligand binding for GPR23. (A) Protocol for GPR23 binding assay. (B) Titration of beads and GPR23 membrane protein to determine optimal mixture ratio. (C) Saturation binding of [3H]LPA to GPR23. (D) Competitive displacement of [3H]LPA by unlabeled LPA. (E) Determination of DMSO tolerance of GPR23 assay. (F) Scatter plot of signals obtained in each well of 384-well plate. All wells in columns 1 through 23 contained 1.25% DMSO and the 16 wells in column 24 contained DMSO and 10 |M LPA before addition of mixed beads and membrane.

types of SPA imaging beads are available including YSi-WGA, polyvinyltoluene-WGA (PVT-WGA) and PVT-WGA treated with polyethyleneimine (PEI) that can help to reduce non-specific binding. Due to their lower density, PVT-WGA beads settle more slowly in solution than YSi-WGA beads and are preferred for automated assays. When we compared the PVT-WGA and PVT-WGA-PEI beads, we observed no difference in the total or non-specific binding and so PVT-WGA beads were used in subsequent work.

To optimize the signal-to-background (S/B) ratio of our assay, we varied systematically the concentrations of membrane proteins and beads (Figure 4.2B). We found that with 20 nM [3H]LPA, total [3H]LPA binding generally increased directly with increasing amounts of both membrane proteins and beads. However, the non-specific binding (in the presence of 10 pM unlabeled LPA) also increased and so the S/B ratio was reduced. Based on these results, we determined that the optimal condition for our GPR23 binding SPA was 20 pL of buffer containing 250 pg of PVT-WGA beads premixed with 2 pg of GPR23 membranes.

We used the GPR23 SPA to determine the equilibrium binding affinity of [3H]LPA for GPR23 receptors. Using the conditions described above, various concentrations of [3H]LPA were added to each well of the plate and incubated at room temperature for 2 hr in the dark prior to reading. Figure 4.2C shows the saturation binding curve for [3H]LPA and GPR23. The specific [3H]LPA binding was calculated by subtracting the non-specific signal in the presence of 10 pM unlabeled LPA from the total signal. Under these assay conditions, the specific binding was concentration-dependent and reached a plateau, consistent with a receptor-mediated mechanism.

We calculated the dissociation constant (Kd) of [3H]LPA to be 21 nM and the receptor density (Bmax) to be 18 pmol/mg protein, indicating a high level of GPR23 expression. Importantly, membranes prepared from uninduced cells displayed no specific [3H]LPA binding, suggesting that the signal we observed with membranes prepared from induced cells was GPR23-dependent. In order to establish an optimized displacement binding assay, it is generally recommended to use the labeled ligand at a concentration close to its Kd value. If the concentration used is much less than Kd, then the ligand may be depleted during the experiment, leading to a rightward shift of the displacement curve and a resulting error in the IC50 value. On the other hand, if the concentration used is much greater than Kd, then not only would the non-specific background signal increase, but there would also be an increase in the amount of radioisotope used and the amount of radioactive waste generated.

One advantage of SPA is that the signal can be measured repeatedly after adding the reagents, allowing the kinetics and stability of the binding reaction to be determined. We found that the specific [3H]LPA binding increased with time, reaching a maximum value 2 hr after radioligand addition. Thereafter, the signal remained stable for another 3 hr before gradually declining (data not shown). Knowledge of receptor binding kinetics and signal stability will facilitate efficient scheduling of a primary screening campaign, during which it is preferable to determine the signal at equilibrium. When the signal remains stable for a long time after reaching equilibrium, larger batches of test plates can be processed via automation, whereas when the signal is short-lived, it is possible to process only smaller batches of plates.

Having established that our assay protocol permitted measurement of the specific binding of [3H]LPA to GPR23, we examined next the ability of unlabeled LPA to displace [3H]LPA from the receptor. Different concentrations of unlabeled LPA (1 pM to 3 pM) were added to each well, followed by 20 pL of premixed beads and membrane. After incubation at room temperature for 30 min, 10 pL of [3H]LPA (final concentration 20 nM) were added and all reagents incubated at room temperature for 2 hr in the dark prior to reading. We found that unlabeled LPA displaced [3H]LPA in a concentration-dependent manner with a calculated inhibitory constant (Ki) of 14 nM (Figure 4.2D).

During pharmacology and HTS experiments, small molecules are typically dissolved in 100% dimethylsulfoxide (DMSO) and so we determined the effect of this substance on assay performance. We found that total binding and non-specific binding were not significantly affected by DMSO up to a final concentration of 2% (Figure 4.2E). This is advantageous because the DMSO tolerance of a GPR23 cell-based assay would likely be much lower than 2% and so small molecules could be studied in the GPR23 binding SPA with less dilution of the original DMSO-based stock solution i.e., at higher test concentrations.

Figure 4.2F shows a scatter plot of the individual signals in the 384 wells of a representative plate in a DMSO trial run (final concentration is 1.25%) that was used to simulate assay performance during a primary screen. The wells in column 23 contained DMSO only, and these represented our high control wells. The total binding signal in these 16 wells was 730 ± 41 counts per minute [c.p.m., mean ± standard deviation (S.D.) of mean]. The wells in column 24 contained 10 pM unlabeled LPA in 1.25% DMSO, and these represented our low control wells. We estimated the non-specific binding signal in these 16 wells to be 163 ± 16 c.p.m. (mean ± S.D.). The wells in columns 1 through 22 contained DMSO only and these can be considered to simulate the behavior of the test wells of a primary screening plate containing inactive compounds. Not surprisingly, the signal in these 352 wells was of similar magnitude to the total signal in the high control column. We calculated that our assay had a S/B ratio of 4.5 and a Z' factor of 0.7, indicating that this GPR23 SPA would be acceptable for use in primary screening (Zhang, Chung, and Oldenburg 1999).

0 0

Post a comment