Chirality of Novel Bitopic Agonists Determines Unique Pharmacology at the Dopamine D3 Receptor

The dopamine D2/D3 receptor (D2R/D3R) agonists are used as therapeutics for Parkinson’s disease (PD) and other motor disorders. Selective targeting of D3R over D2R is attractive because of D3R’s restricted tissue distribution with potentially fewer side-effects and its putative neuroprotective effect. However, the high sequence homology between the D2R and D3R poses a challenge in the development of D3R selective agonists. To address the ligand selectivity, bitopic ligands were designed and synthesized previously based on a potent D3R-preferential agonist PF592,379 as the primary pharmacophore (PP). This PP was attached to various secondary pharmacophores (SPs) using chemically different linkers. Here, we characterize some of these novel bitopic ligands at both D3R and D2R using BRET-based functional assays. The bitopic ligands showed varying differences in potencies and efficacies. In addition, the chirality of the PP was key to conferring improved D3R potency, selectivity, and G protein signaling bias. In particular, compound AB04-88 exhibited significant D3R over D2R selectivity, and G protein bias at D3R. This bias was consistently observed at various time-points ranging from 8 to 46 min. Together, the structure-activity relationships derived from these functional studies reveal unique pharmacology at D3R and support further evaluation of functionally biased D3R agonists for their therapeutic potential.


Introduction
Dopamine (DA) is a major neurotransmitter in the central nervous system responsible for various physiological functions such as motor control, cognition, reward, pain, and memory and learning [1]. Dopamine signaling is mediated by five G protein-coupled receptors (GPCRs) classified into two subgroups based on distinct sequence homologies and signal transduction activities [2]: D 1 -like receptors that primarily couple to Gs protein and enhance the activity of adenylyl cyclase, leading to an increase in intracellular cAMP production, and D 2 -like receptors that primarily couple to the Gi/Go class of G proteins and suppress the activity of adenylyl cyclase. The dopamine D 2 and D 3 receptors (D 2 R and D 3 R), both belonging to the D 2 -like receptor subfamily, represent the major targets for neuropsychiatric disorders such as schizophrenia, Parkinson's disease (PD), and substance use disorders (SUDs) [3,4]. Between the two receptor subtypes, selective targeting of D 3 R would have lower potential for side effects because of its relatively Thus, there is a need for biophysical characterization of novel agonists with D 3 R specific functional selectivity before evaluating whether such agonists can provide therapeutic value in in vivo models and beyond.
Bitopic ligands are comprised of a primary pharmacophore (PP) that binds to the OBS-the endogenous ligand binding site, and a secondary pharmacophore (SP) that binds to a secondary binding pocket (SBP), connected by a chemically defined linker [34][35][36]. Previous studies reveal that the bitopic ligand strategy can provide improved receptor subtype selectivity, affinity, and functional selectivity [37][38][39]. Additionally, bitopic ligands can also confer unique receptor signaling an example of which is the bitopic ligand SB269,652 that behaves as an allosteric antagonist at D 3 Rs and D 2 Rs [40,41]. Indeed, we have successfully utilized a bitopic design strategy to synthesize potent, selective, and G protein biased full agonists at D 2 R [42]. Using a similar strategy, we recently synthesized bitopic D 3 R compounds with increased binding affinity and selectivity for D 3 R demonstrated in radioligand binding studies [43]. In this study, we utilize bioluminescence resonance energy transfer (BRET) based biophysical functional assays to probe structure activity relationships (SAR) in conferring D 3 R over D 2 R functional selectivity.
Specifically, we tested a series of congeneric bitopic compounds with their PP based on PF592,379 ( Figure 1), an agonist at D 3 R developed by Pfizer for the treatment of sexual dysfunction and pain [43][44][45]. Among the series of compounds generated in Battiti et al. [43,46], we selected the ones that showed interesting SAR in radioligand binding assays for functional characterization [43]. In particular, we first selected the compounds that meets at least one of the two criteria: (i) a Ki < 32 nM in D 3 R binding (ii) a D 3 R over D 2 R selectivity > 22 fold. In order to comprehensively probe the role of chirality in conferring unique pharmacology, we further evaluated the corresponding racemic mixture or stereoisomers of the selected compounds in our BRET assays. First, a 3,4-dihydroquinoline-2(1H)-one SP, inspired by the antipsychotic D 2 R/D 3 R partial agonist Aripiprazole, was tethered to the PP with a butyl linker to generate rac-AB04-35, a mixture of two diastereoisomers. Based on the chirality at the PP morpholine ring, specifically in the 2-position, trans-(2R, 5S)-AB04-95 and cis-(2S, 5S)-AB04-96 were prepared via diastereospecific synthesis [43]. Second, a 2-indole amide SP was connected to the PP with the same butyl linker. In order to further assess the role of chirality in the PP, both trans-(2R, 5S)-AB04-87 and cis-(2S, 5S)-AB04-88 diastereoisomers were synthesized. Third, based on the observation of the optimal cis-(2S, 5S)-PP stereochemistry for D 3 R binding, the 2-indole amide SP was connected via a more rigid trans-cyclopropyl [47] containing linker to generate rac-FOB02-04 as a diastereoisomeric mixture of the trans cyclopropyl ring. Chiral resolution of the cyclopropyl linker gave two different isomers (1R, 2S)-FOBO2-04A and (1S, 2R)-FOB02-04B ( Figure 1). Investigating pairs of diastereoisomers, we report functionally selective bitopic compounds that only show the bias characteristic in one stereoisomer, underscoring the importance of stereochemistry as a fundamental structural characteristic in D 3 R functional selectivity. For the bitopic ligand design, the primary pharmacophore (PP) scaffold used was inspired by PF592,379. Several bitopic ligands based on the PF592,379 moiety were synthesized [43], among which the ligands shown were tested for functional characterization.

Bioluminescence Resonance Energy Transfer (BRET) Studies
The BRET-based Go protein activation and β-arrestin recruitment assays were performed as described previously [34,42]. Go protein activation assay uses Renilla luciferase 8 (Rluc8; provided by Dr. S. Gambhir, Stanford University, Stanford, CA, USA)-fused GαoA and mVenus-fused Gγ2 as the BRET pair ( Figure 2a). β-arrestin recruitment assay uses RLuc8-fused D3R or D2R and mVenus-fused β-arrestin2 as the BRET pair ( Figure 2b). HEK293T cells were transiently transfected with the above constructs using polyethyleneimine (PEI) at a ratio of 2:1 (PEI:total DNA by weight). After ~48 h of transfection, cells were washed, harvested, and resuspended in PBS + 0.1% glucose + 200 µM Na bisulfite buffer. Then, 200,000 cells were transferred to each well of the 96-well plates (White Lumitrac 200, Greiner bio-one, Monroe, NC, USA) followed by addition of 5 µM coelenterazine H, a luciferase substrate for BRET. Test compounds, reference D2/D3 agonistquinpirole (Tocris Bioscience, Minneapolis, MN, USA), and vehicle controls were then added by Nimbus liquid handling system (Hamilton, Reno, NV, USA) with its stamping protocol and cells were incubated at 37 °C for 10 min. BRET signal was then measured using a Pherastar FSX plate reader (BMG Labtech, Cary, NC, USA). For kinetic experiments, cells were incubated at 37 °C within the Pherastar FSX plate reader (BMG Labtech, Cary, NC, USA) with BRET signal measurements taken at various time-points ranging from 2-46 min. BRET ratio was calculated as the ratio of mVenus (530 nm) over RLuc8 (480 nm) emission. For the bitopic ligand design, the primary pharmacophore (PP) scaffold used was inspired by PF592,379. Several bitopic ligands based on the PF592,379 moiety were synthesized [43], among which the ligands shown were tested for functional characterization.

Bioluminescence Resonance Energy Transfer (BRET) Studies
The BRET-based Go protein activation and β-arrestin recruitment assays were performed as described previously [34,42]. Go protein activation assay uses Renilla luciferase 8 (Rluc8; provided by Dr. S. Gambhir, Stanford University, Stanford, CA, USA)-fused Gα oA and mVenus-fused Gγ 2 as the BRET pair ( Figure 2a). β-arrestin recruitment assay uses RLuc8fused D 3 R or D 2 R and mVenus-fused β-arrestin2 as the BRET pair ( Figure 2b). HEK293T cells were transiently transfected with the above constructs using polyethyleneimine (PEI) at a ratio of 2:1 (PEI:total DNA by weight). After~48 h of transfection, cells were washed, harvested, and resuspended in PBS + 0.1% glucose + 200 µM Na bisulfite buffer. Then, 200,000 cells were transferred to each well of the 96-well plates (White Lumitrac 200, Greiner bio-one, Monroe, NC, USA) followed by addition of 5 µM coelenterazine H, a luciferase substrate for BRET. Test compounds, reference D 2 /D 3 agonist-quinpirole (Tocris Bioscience, Minneapolis, MN, USA), and vehicle controls were then added by Nimbus liquid handling system (Hamilton, Reno, NV, USA) with its stamping protocol and cells were incubated at 37 • C for 10 min. BRET signal was then measured using a Pherastar FSX plate reader (BMG Labtech, Cary, NC, USA). For kinetic experiments, cells were incubated at 37 • C within the Pherastar FSX plate reader (BMG Labtech, Cary, NC, USA) with BRET signal measurements taken at various time-points ranging from 2-46 min. BRET ratio was calculated as the ratio of mVenus (530 nm) over RLuc8 (480 nm) emission. ules 2021, 11, x 5 of 16  Data were collected from at least 3 independent experiments and normalized to maximal response by quinpirole as 100% and response by vehicle as 0%. Concentration response curves (CRCs) were generated using a non-linear sigmoidal dose-response analyses using Prism 8 (GraphPad Software, San Diego, CA, USA) and presented as mean ± SEM.

Bias Factor Analysis
To evaluate whether the test compounds exhibited G protein versus β-arrestin signaling bias, bias factors were calculated using the method as follows [1]: An arbitrary but stringent cut-off of ≥ ±2.0 (in logarithmic scale) was chosen to identify biased ligands. Bias factor values >2.0 represent bias towards G protein while values below <−2.0 represent bias towards β-arrestin.

An In-House Program for Kinetics Analysis of Functional Assay Data
In this study, the BRET signals for each 96-well plate were detected every two minutes using the BMG Pherastar FSX plate reader (BMG Labtech, Cary, NC, USA), resulting in 23 datasets for a 46 min measurement in one raw data file per plate. Such amount of data is beyond what can be conveniently and reliably processed by manual extraction, transformation, and normalization, before the regression analysis by Prism. Thus, we developed an in-house python program that can process and analyze the kinetics of functional assay data. While this program was configurated to fit the raw file output format of the plate reader used in this study, it can be easily adapted to process other output formats, time intervals, and plate maps (i.e., how the rows and columns of 96-well plates are configured for the dose-response measurements). This program also has the capability to process multiple files, based on predefined file locations in a configuration file. In this configuration file, for each raw file to be processed, it also includes time intervals, receptor construct, test compounds, and concentration ranges for each plate.
Each raw BRET data file was first preprocessed by detecting the data set for each time cycle. The response values were calculated as the ratio between 475-30B and 535-30A data for each well of the 96-well plate in the same time cycle. We took the average of response values for each compound at each concentration in the same time cycle. We then fitted response values to the sigmoidal dose-response function.
Sigmoidal dose-response function: S(x) = top−bottom 1+10 (x−logEC50) + bottom, top and bottom are the maximum and minimum of the response values, respectively, x is the logarithm of the concentration, and logEC 50 is the x value when the response is halfway between bottom and top.
We used the scipy.optimize.curve_fit module (version 1.5.2) [48] to perform this fitting process. In this curve fit module, we chose 'lm' as the optimization method type, which can replicate the regression result using Prism 8 (GraphPad Software, San Diego, CA, USA). The initial guess values were the minimum response value for "bottom", the maximum response value for "top", and the halfway value of the log(concentration) range for "logEC 50 ". After the fitting, the E max was calculated as the difference between the optimized "bottom" and the optimized "top", and the log (concentration) resulting in half of E max is the logEC 50 .
To demonstrate functional kinetics, the program integrates the regression results at each time point and plots the E max and logEC 50 evolutions along the time ( Figure 4).

Statistics
Statistical significance values were calculated using Prism 8 (GraphPad Software, San Diego, CA, USA)'s ordinary One-way ANOVA (independent variable: compound treatment, dependent variables: efficacy or pEC 50 s) followed by Dunnett's multiple comparisons tests. For kinetic data, statistical significance values were calculated using GraphPad Prism's ordinary two-way ANOVA (independent variables: compound treatment and time-point, dependent variables: efficacy or pEC 50 s) followed by Sidak's multiple comparisons tests. For the above, two multiple comparisons, '*' represents a significance of p < 0.05; '**' of p < 0.01; '***' of p < 0.001 and '****' of p < 0.0001 compared to quinpirole. Dunnett's multiple comparisons tests were also performed against AB04-87 with 'δ' representing significance of p < 0.05; 'δδ' of p < 0.01; 'δδδ' of p < 0.001 and 'δδδδ' of p < 0.0001 compared to AB04-87. Data are reported from more than three experiments performed in triplicate. In the case where the data points could not be fitted into the non-linear sigmoidal dose-response equation, the pharmacological parameters are reported as ND (not determined).

Bitopic Compounds Exhibit Varying Pharmacological Profiles at Both D 3 R and D 2 R Compared to the Reference D 2 R/D 3 R Agonist Quinpirole
The bitopic compounds characterized in this study all have the PF592,379 scaffold as the primary pharmacophore (PP). For the chiral center at the 2-position of the morpholine ring in this scaffold, while AB04-88, rac-FOB02-04, FOB02-04A, FOB02-04B, and AB04-96 are in the cis conformation, AB04-95 and AB04-87 possess trans stereochemistry. rac-AB04-35 is the diastereomeric mixture of the AB04-95 and AB04-96, which have the same butyl linker and 3,4-dihydroquinoline-2(1H)-one as the SP (Figure 1). While AB04-87 and AB04-88 also have the same butyl linker, they differ from these compounds in their SP (2-indole amide). rac-FOB02-04, FOB02-04A, and FOB02-04B have the same 2-indole amide SP but with different trans-cyclopropyl containing linker. To investigate the impact of these differences in stereochemistry, linker, and SP on their functional properties, we evaluated the pharmacological profiles of these bitopic compounds with BRET-based Go protein activation and β-arrestin recruitment assays (see Section 2).
The cis-isomers of bitopic ligands are more potent at D 3 R than their trans-isomers. AB04-96 and AB04-95 only differ in the chirality at the 2-position of the morpholine ring in their PP, i.e., in cisand trans-stereochemistry, respectively (Figure 1) [43]. The same difference in stereochemistry is seen between AB04-88 and AB04-87, another diastereoisomeric pair. In all four compounds, the same butyl linker was used to connect the PP with their corresponding SPs. Given that chirality can considerably modulate pharmacological profiles [38,49], we compared these two pairs of ligands in their ability to promote D 3 R Go protein activation and β-arrestin recruitment.
As shown in Figure 2e-h and Table 1, in both cases, the cis-stereochemistry (AB04-88 and AB04-96) of the morpholine ring confer higher potencies as well as efficacies compared to their corresponding trans diastereoisomers (AB04-87 and AB04-95, respectively). The cis-AB04-88 exhibited a~2188-fold increase in potency (p < 0.0001, Dunnett's multiple comparisons test) and an increase in E max of~51.3% (p < 0.0001) compared to its trans-AB04-87 in D 3 R Go protein activation (using the efficacy of quinpirole as 100%) (Figure 2e and Table 1).
In D 3 R mediated β-arrestin recruitment, the cis-AB04-88 exhibited a significant increase in E max of~42.8% (p < 0.0001) and a 17-fold increase in potency compared to its trans-AB04-87 ( Figure 2f and Table 1). Interestingly, while the trans-AB04-95 exhibited a near complete abolishment of D 3 R mediated Go protein activation as well as β-arrestin recruitment, cis-AB04-96 exhibited a robust activity in both assays as shown in Figure 2g-h and Table 1. Thus, the potency and efficacy of AB04-95 cannot be quantitatively determined.

The cis-AB04-88 But Not the trans-AB04-87 Shows D3R Selective Go Protein Bias.
Previous studies have shown that bitopic ligands can not only improve subtype potency and selectivity but can also promote functional selectivity [42,52]. There is an increasing interest in understanding G protein versus β-arrestin biased agonism at GPCRs because of potential therapeutic utility upon activating one pathway over the other. To evaluate whether the compounds shown in Figure 1 can promote D3R specific G protein versus β-arrestin functional selectivity, bias factors were derived using the model that incorporates both efficacy and potency differences into a single index [1].
The biased activity equation used to evaluate G protein versus β-arrestin bias is described in the Section 2. An arbitrary but stringent cut-off of > ± 2.0 (in logarithmic scale) was chosen to identify biased ligands, where values > 2.0 represent bias towards G protein AB04-96 is the bitopic ligand with the same cis-stereochemistry of PP and butyl linker as AB04-88 but with a different SP. Interestingly, this cis-isomer only exhibited a 2.8-fold selectivity (Figure 3d and Table 3) with its corresponding trans-isomer showing near abolishment of activity at both D 3 R and D 2 R (Figure 3c and Tables 1 and 2).

The cis-AB04-88 But Not the trans-AB04-87 Shows D 3 R Selective Go Protein Bias
Previous studies have shown that bitopic ligands can not only improve subtype potency and selectivity but can also promote functional selectivity [42,52]. There is an increasing interest in understanding G protein versus β-arrestin biased agonism at GPCRs because of potential therapeutic utility upon activating one pathway over the other. To evaluate whether the compounds shown in Figure 1 can promote D 3 R specific G protein versus β-arrestin functional selectivity, bias factors were derived using the model that incorporates both efficacy and potency differences into a single index [1].
The biased activity equation used to evaluate G protein versus β-arrestin bias is described in the Section 2. An arbitrary but stringent cut-off of > ± 2.0 (in logarithmic scale) was chosen to identify biased ligands, where values > 2.0 represent bias towards G protein while values < −2.0 represent bias towards β-arrestin. The calculated biased factors are presented in Table 3.
Among all the ligands tested for bias at both D 3 R and D 2 R (Table 3), only the cis-isomer AB04-88 exhibited a biased factor equal to 2.3 ( Figure 3b and Table 3) at D 3 R, suggesting that this ligand promotes D 3 R mediated Go protein bias. Upon inversion at the chiral center in the 2-position to generate trans-isomer AB04-87, the D 3 R mediated Go protein bias is completely lost (Figure 3a and Table 3). A similar trend can be observed at D 2 R, whereby the bias factor for AB04-88 was calculated to be 1.5 whereas the bias factor for AB04-87 could not be quantified (Figure 3a-b and Table 3), suggesting that AB04-88 exhibits a D 3 R selective Go protein bias.
In case of the second set of cis-trans pairs, the bias factor for the cis-AB04-96 was calculated to be 1.0 and 1.1 for D 3 R and D 2 R, respectively. However, the bias factors for its corresponding trans-isomer AB04-95 could not be quantified at both D 3 R and D 2 R because its activities were nearly abolished for both Go protein activation and β-arrestin recruitment. Although AB04-88 and AB04-96 were designed using the same cis form of the PP and the same butyl linker, they differ in their SPs. The observed D 3 R mediated Go protein bias using AB04-88 but not AB04-96 further stresses the role of SPs in conferring unique pharmacological profiles.

Time-Dependent Pharmacological Analysis Reveals Higher Go Protein Activation E max for AB04-88 at Later Time Points
Time-dependent pharmacological profile, often called kinetic context, can be determined by the ligand-binding and receptor-signaling kinetics collectively and can impact biased agonism and its quantification [53,54]. Since our bitopic agonist AB04-88 exhibited D 3 R specific bias towards Go protein, we evaluated the kinetic profiles of AB04-88 in comparison to that of quinpirole. We assessed the kinetic profiles of both efficacies and potencies that constitute the two components used in our D 3 R biased agonism quantification [1], based on the data points collected every 2 min up to 46 min (see Section 2).
As shown in Figure 4, there were no significant differences between the AB04-88 and quinpirole efficacies at D 3 R until the 36 min time-point. Interestingly, after the 36 min time-point, AB04-88 exhibits significantly higher efficacies compared to quinpirole (in all cases p < 0.05; two-way ANOVA followed by Sidak's multiple comparisons test) with maximal increases of > 40% observed at 44 and 46-min time-points ( Figure 4a and Table 4). In contrast, there were no statistically significant differences observed between AB04-88 and quinpirole for EC 50 of Go protein activation (Figure 4c and Table 4) or the efficacy component for β-arrestin recruitment (Figure 4b and Table 4). In contrast, in the β-arrestin recruitment assay, the E max profiles for both AB04-88 and quinpirole nearly overlap up to 36 min, with AB04-88 showing slightly decreased Emax afterwards. The potency of AB04-88 remains significantly and consistently lower compared to quinpirole at all time-points ranging from 4-46 min (in all cases p < 0.05; two-way ANOVA followed by Sidak's multiple comparisons test) (Figure 4d and Table 4). We also calculated the bias factors at all time-points, AB04-88 exhibited a bias at timepoints ranging from 8-46 min (> ±2.0) but without major changes.
These data collectively suggest that, in the case of AB04-88, kinetics analysis revealed a significant Emax increase for Go protein activation at later time points, which, however, did not significantly affect the bias factor.  Statistical significances were calculated using GraphPad Prism's ordinary two-way ANOVA (in each case, **** p < 0.0001) followed by Sidak's multiple comparisons tests with '*' representing significance of p < 0.05; '**' of p < 0.01 and '****' of p < 0.0001 compared to that of quinpirole. We also calculated the bias factors at all time-points, AB04-88 exhibited a bias at timepoints ranging from 8-46 min (> ±2.0) but without major changes.
These data collectively suggest that, in the case of AB04-88, kinetics analysis revealed a significant E max increase for Go protein activation at later time points, which, however, did not significantly affect the bias factor.

Discussion
The discovery of D 3 R selective agonists is of particular interest because such compounds would help ascertain D 3 R specific physiological and pathological functions from that of D 2 R in neurodegenerative disorders such as PD [55]. Furthermore, such agonists may exhibit improved clinical utility in the treatment of such disorders. The bitopic D 3 R ligands evaluated in this study exhibit varying pharmacological profiles compared to the reference D 2 R/D 3 R full agonist quinpirole. The PPs of the bitopic ligands in this study are based on the D 3 R-preferential agonist, PF592,379 and are either a mixture of diastereoisomers or one of the two cisand transisomers based on the chirality of its morpholine ring. In addition, the bitopic ligands differ in their SPs, linkers, and linker stereochemistry.
Comparing the diastereomer pairs within the same ligand structures, the (2S, 5S) enantiomer of the PP morpholine ring shows more efficacious and potent D 3 R G protein signaling. This observation confirms the significance of enantiospecificity in the morpholine ring in achieving D 3 R activation efficacy.
In D 3 R Go protein activation BRET, cis-(2S, 5S)-AB04-88 presenting a 2-indole amide SP connected by a butyl linker, showed a 123-fold selectivity for D 3 R over D 2 R. Interestingly, another compound AB04-96 that has the same PP and linker but with a 3,4-dihydroquinolin-2(1H)-one SP only shows a 2.8-fold D 3 R over D 2 R selectivity. This observation highlights the importance of the SP in promoting subtype selectivity where the 2-indole amide, but not 3,4-dihydroquinolin-2(1H)-one promotes D 3 R subtype selectivity. This is consistent with previous findings where the presence of a 3,4-dihydroquinolin-2(1H)-one moiety improves D 2 R subtype binding [42,56,57]. Unlike the previous radioligand binding study in which FOB02-04A showed a 47.5-fold D 3 R over D 2 R selectivity, we only observed an eight-fold selectivity in Go protein activation BRET. The binding affinities and selectivities observed in radioligand binding studies may not always correlate with cell-based functional assays. The observed discrepancies in these two types of assays may be explained by inherent differences of probe dependency and specific assay conditions [58].
Given the increasing evidence supporting therapeutic utility of functional selectivity in GPCRs, we evaluated the D 3 R bitopic agonists for their signaling bias. AB04-88, which showed the highest potency in D 3 R Go protein activation among all bitopic ligands studied, also exhibited a Go protein activation over β-arrestin recruitment bias by two log units. While the (2S, 5S) cis stereoisomer of the PP appears favorable for Go protein activation over β-arrestin recruitment, AB04-96 with the same cis conformation of PP and linker, but a different SP, does not exhibit functional bias. Similar to the D 3 R over D 2 R selectivity, this difference highlights the significance of the SP in conferring functional bias as well. Future computational studies with the two diastereoisomeric (based on PP morpholine ring) sets of D 3 R bitopic agonists will provide clues as to the interplay between chirality and SP chemistry to provide unique pharmacology.
Increasing evidence suggests the influence of ligand binding kinetics and receptor signaling in biased agonism [53,54]. Since AB04-88 exhibited D 3 R specific Go protein bias at 10 min, we investigated whether this bias would change in a time-dependent manner. The only kinetic profile affected was that of the efficacy of AB04-88 in D 3 R Go protein activation, where it significantly increased compared to quinpirole after 36 min. However, the bias factor was not affected due to the largely stable potencies in both Go protein activation and β-arrestin assays, the difference of which would be the dominant component in the factor when AB04-88 remained a strong partial agonist at the later time-points of β-arrestin recruitment.
Together, our newly developed D 3 R bitopic ligands provide novel tools to further probe the unique pharmacology, specific cellular signaling, and therapeutic potential of D 3 R.