Dopamine D4 Receptor-Selective Compounds Reveal Structure-Activity Relationships that Engender Agonist Efficacy.

The dopamine D4 receptor (D4R) plays important roles in cognition, attention, and decision making. Novel D4R-selective ligands have promise in medication development for neuropsychiatric conditions, including Alzheimer's disease and substance use disorders. To identify new D4R-selective ligands, and to understand the molecular determinants of agonist efficacy at D4R, we report a series of eighteen novel ligands based on the classical D4R agonist A-412997 (1, 2-(4-(pyridin-2-yl)piperidin-1-yl)- N-( m-tolyl)acetamide). Compounds were profiled using radioligand binding displacement assays, β-arrestin recruitment assays, cyclic AMP inhibition assays, and molecular dynamics computational modeling. We identified several novel D4R-selective ( Ki ≤ 4.3 nM and >100-fold vs other D2-like receptors) compounds with diverse partial agonist and antagonist profiles, falling into three structural groups. These compounds highlight receptor-ligand interactions that control efficacy at D2-like receptors and may provide insights into targeted drug discovery, leading to a better understanding of the role of D4Rs in neuropsychiatric disorders.

Introduction

The dopamine D4 receptor (D4R) is a G protein-coupled receptor and a member of the D2-like subfamily of dopamine receptors (including D2R, D3R, and D4R). D2-like receptors have high sequence homology and share a Gαi/o-coupled signaling mechanism, but differ substantially in localization within the brain and at the subcellular level.[1] Compared with D2Rs and D3Rs, D4Rs have the lowest level of expression in the brain and show a unique distribution pattern, with most located in the prefrontal cortex (PFC) and hippocampus. The other D2-like receptors are primarily in the striatum, basal ganglia, and pituitary gland regions, regions associated with D2R-targeting antipsychotic drugs and the motor and endocrine side effects commonly observed with them.[2,3] In contrast, D4Rs expressed in PFC and hippocampus affect attention, exploratory behavior,[3] and performance in novel object recognition[4,5] and inhibitory avoidance[6] cognitive tasks. Therefore, pharmacological activation of D4Rs may be useful to treat cognitive deficits associated with schizophrenia[7−10] and attention-deficit/hyperactivity disorder.[10,11] Additional research has explored D4R agonism as a strategy to reduce the adverse effects of opioid drugs like morphine.[12,13] D4R antagonism may be useful to treat substance use disorders (SUDs), particularly psychostimulant addiction, and l-DOPA-induced dyskinesias.[10,14−20] The importance of targeting D4Rs in treating these complex pathologies, especially with regards to the extent of receptor activation or inhibition, remains unknown, partially because of a lack of suitable compounds for investigating these pathways.

A-412997 (1, 2-(4-(pyridin-2-yl)piperidin-1-yl)-N-(m-tolyl)acetamide, Figure ) was initially characterized as a “full agonist” (83% intrinsic activity) at D4R, with high selectivity over D2R and D3R and in vivo effects that included induction of penile erection in rats.[21,22] Subsequent in vivo evaluations showed improved cognitive performance in social recognition tasks, novel object recognition tasks, and 5-trial repeated acquisition inhibitory avoidance tasks following treatment by 1 [or similar D4R agonists PD168077 (2) and CP226269 (3)], suggesting an important role for D4R signaling in mediating short-term memory and cognition.[5,23]

Figure 1

Three classic D4R-selective partial agonists.

The goals of this study were to develop new D4R agonists with a range of efficacy levels and to identify the molecular components that engender ligand efficacy at D4R. To that end, we employed a rational drug design strategy incorporating classic structure–activity relationship (SAR) analysis around lead compound 1. These studies were enhanced by detailed in silico molecular dynamics (MD) simulations exploiting the recently reported crystal structure of D4R.[24] Furthermore, comparative analyses were done using the D3R crystal structure[25] and the recently reported D2R structure.[26]

We synthesized a library of analogues primarily featuring modifications in the phenylpiperidinyl region of 1, with additional variations in linker chain length and substitutions on the amidylphenyl region. Following extensive in vitro analyses, including binding and functional studies, we determined that selected modifications resulted in novel analogues with improved subtype selectivity. Furthermore, we identified three classes of modifications that resulted in altered efficacy profiles at all D2-like receptors. In order to determine key receptor–ligand interactions, and identify the molecular substrates of a putative “efficacy switch,” the library was docked in receptor models of D2R, D3R, and D4R using MD simulations.

Chemistry

Ligands were synthesized as outlined in Scheme using routine N-alkylation reactions previously reported.[21,27] The commercially available m-toluidine 4 was converted to intermediate 2-chloro-N-(m-tolyl)acetamide 5 by reacting with 2-chloroacetyl chloride in the presence of triethylamine and ethyl acetate at room temperature.[28] Using the same procedure, intermediates 14, 19, and 24 were synthesized in a similar manner, as indicated in Scheme , with either a one- or two-carbon linker. The intermediate compounds 5, 14, 19, and 24 were used to alkylate different commercially available arylpiperazine or arylpiperidine amines in the presence of K2CO3 in CH3CN under reflux conditions to yield the desired target compounds 6–9, 10–13, 15–17, 20–22, and 25–28, respectively, with the exception of the synthesis of 1-(naphthalen-1-yl)piperazine which was previously reported[29] via nucleophilic substitution reaction with naphthalen-1-amine.

Scheme 1

Synthesis of 2-(4-(Pyridin-2-yl)piperidin-1-yl)-N-(m-tolyl)acetamide Analogues

Reagents and conditions: (a) triethylamine, EtOAc, RT; (b) CH3CN, K2CO3, reflux, appropriate arylpiperazine or arylpiperidine.

Pharmacological Results and Discussion

SARs at Dopamine D2-like Receptors

A primary objective of this study was to design ligands with high D4R binding affinity and subtype selectivity. The compound 1 and several designed analogs are shown in Figure . In order to obtain D4R ligands with high affinity and selectivity, using compound 1 as our lead compound, we employed three modification strategies, creating 2-(piperidin-4-yl)pyridinyl analogs, altering the linker chain length, and creating N-(m-tolyl)acetamide analogs.

Figure 2

Three classes of modifications to the structure of 1 resulting in differing binding and efficacy profiles at D2-like receptors. (A) Substitution of the piperidine ring for piperazine induced a gain of efficacy at D2R and D3R with insubstantial changes to D4R efficacy. (B) Substitution of the pyridine ring with a phenyl or napthyl moiety produced modest D4R subtype selectivity improvements and lowered partial agonist efficacy at D4R with no agonist activity at D2R or D3R. (C) Para-substituted pyridine rings produced highly D4R-selective antagonists.

Of note, when 1 was evaluated in two different functional assays, its profile was clearly that of a partial agonist rather than a full agonist as it is often described in the literature. In the agonist mode for both the cyclic AMP (cAMP) accumulation and β-arrestin recruitment assays, 1 had an Emax of 61.9 and 22.5%, respectively, when normalized to dopamine.

The 2-pyridine moiety of 1 was replaced with a phenyl in 6, para-tolyl in 7, 4-chlorophenyl in 8, and 5-methylpyridin-2-yl in 9. The piperidine attached to the linker chain was replaced with a piperazine to form 10, replaced with a pyrimidine to form 11, replaced with a 5-chloropyridin-2-yl to form 12, and replaced with a naphthyl substituent to obtain 13. To evaluate the contribution of the alkyl chain to the binding affinity and selectivity, we synthesized alkyl chain length analogs of compounds 1, 10, and 11, adding an extra methylene to the linker chain in compounds 15, 16, and 17, respectively. Finally, we probed the contribution of the N-(3-methylphenyl)acetamide moiety via replacement of the methyl with ethyl (compounds 20, 21, and 22, compared to compounds 1, 10, and 12, respectively) or replaced of the entire N-(3-methylphenyl)acetamide moiety with heteroaromatics (compounds 25–28).

In order to best evaluate comparative affinities, two different radioligands were used in competition binding studies: [3H]N-methylspiperone, a high-affinity D2-like antagonist, and [3H]-(R)-(+)-7-OH-DPAT, a D2-like agonist. Importantly, the binding affinities of D2-like agonists and high-efficacy partial agonists are considerably higher when competing against an agonist radioligand because high-affinity agonist binding incorporates an efficacy measure in that the greater the efficacy for inducing G protein coupling, the greater the “apparent” affinity will be. On the other hand, antagonist binding, and competition for it, is unlinked from efficacy and therefore unbiased. Therefore, because these radioligands probe different receptor states, they provide complimentary views of ligand binding,[30] which are particularly valuable when examining affinity of partial agonists.

Several modifications of 1 resulted in modest improvements in D4R affinity as measured by competition assays with [3H]N-methylspiperone (up to ∼3-fold) and [3H]-(R)-(+)-7-OH-DPAT (up to ∼3-fold). However, marked improvements in D4R selectivity over D2R and D3R resulted from a variety of modifications, typically driven by a loss of affinity at D2R and D3R.

2-Pyridine substitutions resulted in a potency gain when the piperidinyl moiety was replaced with piperazinyl (e.g., 10 and 21). Adding an extra methylene to the linker chain, as in compounds 15, 16, and 17, significantly diminished D4R affinity and selectivity. These results are consistent with previous studies that determined the importance of carboxamide linker length for D2-like receptor selectivity.[31] Replacement of the methyl with an ethyl at the N-(3-methylphenyl)acetamide moiety (compounds 20, 21, and 22) did not substantially alter affinity or selectivity for D4R compared to methyl analogues 1, 10, and 12, respectively. Replacement of the entire N-(3-methylphenyl)acetamide moiety with heteroaromatics (compounds 25–28) uniformly led to loss of affinity and selectivity.

Overall, we noted three broader classes of modifications with distinct binding and efficacy profiles across the D2-like receptors; as outlined in Figure , these include (1) substitution of the piperidine ring for piperazine, (2) substitution of the pyridine ring with a phenyl or napthyl moiety, and (3) para-substituted pyridine rings. These classes formed the basis for further SAR profiling and modeling studies using MD simulations.

The parent compound, 1, showed 115-fold and 31-fold higher affinity for D4R over D2R and D3R, respectively, as measured by [3H]N-methylspiperone competition. When examined using [3H]-(R)-(+)-7-OH-DPAT competition, 1 had higher affinity at all subtypes (consistent with an agonist radioligand being displaced by a compound that favors the activated receptor[30]) and showed a similar selectivity profile of 64-fold and 42-fold higher affinity for D4R over D2R and D3R, respectively. Full binding results are presented in Table . Functional characterization revealed 1 to be a partial agonist at D4R as measured in β-arrestin assays (Emax = 22.5%, EC50 = 473 nM) (Figure A,B) and cAMP inhibition assays (Emax = 61.9%, EC50 = 2.7 nM) (Figure B,C). The higher efficacy observed in the cAMP assay is likely due to spare receptors and/or amplification of cAMP accumulation versus recruitment of β-arrestin. Consistent with a partial agonist profile, 1 and related analogs were partial antagonists when run in antagonist mode (Figure B,D), blocking function to a similar degree as their maximal agonist activity. This would be expected for a compound that is a partial agonist that maintains affinity for the orthosteric part of the receptor, thereby acting as a partial antagonist in antagonist assays. Importantly, 1 showed no measurable agonist response on D2R-mediated β-arrestin recruitment but behaved as a low affinity full antagonist (Figure E). Furthermore, 1 has very low potency and efficacy at the D3R (Figure F). Complete functional results are presented in Tables and 3. These data indicate that 1 is a potent and highly selective partial agonist at the D4R.

Table 1

Human Dopamine D2-like Receptor Binding Data in HEK293 Membranes for Ligands with Varying Arylpiperazine and Arylamide Moietiesa

Ki values determined by competitive inhibition of [3H]N-methylspiperone or [3H]-(R)-(+)-7-OH-DPAT binding in membranes harvested from HEK293 cells stably expressing hD2R, hD3R, or hD4R. All Ki values are presented as means ± SEM.

Figure 3

Compounds 10 (red) and 21 (gray) show similar pharmacology to parent compound 1 (black). D4R-expressing stable cells lines were plated and compounds were assayed for agonist (A) and antagonist (B) activity on β-arrestin recruitment. Similarly, D4R-mediated inhibition of cAMP accumulation was also examined in both agonist (C), and antagonist (D) modes, as indicated. Assays were conducted as described in the Experimental Methods; briefly, agonist assays were conducted by incubating the cells with the indicated concentration of test compound and measuring luminescence. Antagonist assays were conducted by incubating the compound with an EC80 concentration of dopamine (1 μM for β-arrestin and 10 nM in cAMP) and the indicated concentration of the test compound. For cAMP assays, cells were first stimulated with 10 μM forskolin. Agonist mode assays are expressed as a percentage of the maximum dopamine response, whereas antagonist mode assays are expressed as a percentage of dopamine’s EC80 response. Emax and EC50 values are shown in Tables and 3. Data were fit using nonlinear regression of individual experiments performed in triplicate and are shown as means ± SEM; n = 3. Dopamine and sulpiride were run during each assay as positive controls for a full agonist and full antagonist respectively (data not shown). Compounds were also tested for both agonist and antagonist activity on cells stably expressing the closely related D2R (E) or D3R (F). Assays were conducted as described in the Experimental Methods. Agonist mode assays (open symbols) are expressed as a percentage of the maximum dopamine response observed for each receptor, whereas antagonist mode assays (solid symbols) are expressed as a percentage of dopamine’s EC80 response. Emax and EC50 values are shown in Tables and 3. Data were fit using nonlinear regression of individual experiments performed in triplicate and are shown as means ± SEM; n = 3.

Table 2

Efficacy as Measured via Modulation of cAMP Accumulationa

	D2R efficacy				D4R efficacy				EC50	IC50
compound	cAMP Emax %b	cAMP EC50(nM)	cAMP Ant. %c	cAMP IC50(nM)	cAMP Emax %b	cAMP EC50(nM)	cAMP Ant. %c	cAMP IC50(nM)	D2R/D4R	D2R/D4R
1	inactive	inactive	ND	>50000	61.9 ± 4.7	2.7 + 0.9	53.8 ± 6.0	68.4 ± 32.8	ND	>735
6	inactive	inactive	100 + 0.00	16447 + 3540	32.9 ± 3.9	15.4 ± 13.2	46.7 ± 6.0	2.0 ± 0.05	ND	8224
7	inactive	inactive	100 ± 0	44834 ± 28125	inactive	inactive	95.8 ± 2.2	3064 ± 1220	ND	15
8	inactive	inactive	97.5 ± 2.5	71437 ± 28563	inactive	inactive	100 ± 0	70157 ± 20766	ND	1.0
9	inactive	inactive	100 ± 0	71065 ± 20 585	inactive	inactive	100 ± 0	453 ± 15	ND	157
10	18.96 ± 5.2	763 ± 386	ND	>100000	64.2 ± 5.7	3.6 ± 1.3	43.2 ± 1.8	82.7 ± 37.9	214	>1210
11	54.4 ± 7.5	2092 ± 46	ND	>100000	64.6 ± 4.2	3.4 + 2.0	45.0 ± 7.7	463 ± 157	612	>216
15	83.1 ± 4.2	50.1 ± 25	ND	>100000	28.1 ± 1.6	349 ± 75	77.6 ± 5.2	6343 ± 2524	0.14	>16
16	79.7 ± 8.4	154 ± 31	ND	>100000	30.0 ± 2.1	612 ± 563	84.2 ± 6.0	1629 ± 255	0.25	>61
17	inactive	inactive	ND	>100000	13.7 ± 1.2	568 ± 456	87.0 ± 3.4	2120 ± 534	ND	>47
12	inactive	inactive	98.3 ± 1.7	66077 ± 18646	inactive	inactive	93.4 ± 2.6	4701 ± 1466	ND	14
13	inactive	inactive	100 ± 0	68329 ± 31671	27.8 ± 8.4	108.5 ± 94.3	73.8 ± 13.6	2521 ± 1067	ND	27
20	inactive	inactive	96.6 ± 3.5	16278 ± 11601	25.6 ± 7.2	539 ± 151	70.8 ± 15	1908 ± 242	ND	9
21	18.80 ± 8.19	1600 ± 396	88 ± 6.1	40466 ± 29968	58.0 ± 1.8	28.7 ± 9.9	58.4 ± 9.7	1311 ± 814	56	31
22	inactive	inactive	100 ± 0	46795 ± 27644	inactive	inactive	100 ± 0	7059 ± 1136	ND	7
25	38.6 ± 3	1965 ± 44	100 ± 0	>100000	47.3 ± 7.9	1075 ± 390	100 ± 0	86493 ± 3130	2.0	>1.1
26	inactive	inactive	100 ± 0	86617 ± 13383	inactive	inactive	100 ± 0	40000 ± 9421	ND	2
27	inactive	inactive	98.1 ± 1.5	94255 ± 5745	inactive	inactive	100 ± 0	>100000	ND	<1
28	inactive	inactive	76.2 ± 17.6	72516 ± 18052	inactive	inactive	100 ± 0	>100000	ND	<1

Values determined by nonlinear regression of individual experiments run in triplicate as detailed in materials and methods under cAMP accumulation assays. All EC50, IC50, and Emax values are presented as means ± SEM; n = 3–4. ND indicates not determined due to an incomplete curve. Inactive indicates no measurable activity in indicated assay.

A measure of agonism as defined by the maximum inhibition of cAMP observed for each compound.

A measure of antagonism as defined by the maximum blockade of dopamine mediated cAMP inhibition by each compound.

Table 3

Efficacy as Measured via Modulation of β-Arrestin Recruitmenta

	D2R efficacy				D3R efficacy				D4R efficacy				EC50		IC50
compound	β-arr Emax	β-arr EC50(nM)	β-arr Ant. %	β-arr IC50	β-arr Emax	β-arr EC50(nM)	β-arr Ant. %	β-arr IC50	β-arr Emax	β-arr EC50(nM)	β-arr Ant. %	β-arr IC50	D2R/D4R	D3R/D4R	D2R/D4R	D3R/D4R
1	inactive	inactive	94.8 ± 2.8	5846 ± 1802	ND	>100000	ND	>100000	22.5 ± 3.98	473 ± 457	81.7 ± 2.7	191 ± 98	ND	>4	31	>524
6	inactive	inactive	99.7 ± 0.3	7692 ± 2301	inactive	inactive	ND	>50000	14 ± 0.3	242 ± 89	93.3 ± 1.8	135 ± 65	ND	ND	57	>371
7	inactive	inactive	100 ± 0	89153 ± 10847	inactive	inactive	inactive	inactive	inactive	inactive	100 ± 0	7352 ± 1749	ND	ND	12	ND
8	inactive	inactive	100 ± 0	>100000	inactive	inactive	inactive	inactive	inactive	inactive	100 ± 0	42357 ± 30018	ND	ND	>2	ND
9	inactive	inactive	100 ± 0	>100000	inactive	inactive	inactive	inactive	inactive	inactive	98.1 ± 2	394 ± 78	ND	ND	>254	ND
10	23.9 ± 5.1	26175 ± 12448	76.5 ± 6.9	16010 ± 5174	49.4 ± 2.0	6354 ± 2617	inactive	>100000	30.7 ± 6.4	394 ± 294	78.9 ± 3.1	313 ± 215	66	>16	51	>320
11	39.6 ± 5.3	8321 ± 3455	78.9 ± 8.6	25008 ± 5017	58.4 ± 6.6	5581 ± 1614	inactive	>100000	24.7 ± 5	278 ± 167	80.6 ± 3.0	197 ± 115	30	>20	127	>509
15	48.9 ± 6.8	2480 ± 1834	72.1 ± 7.9	7627 ± 1573	55.4 ± 1.57	970 ± 115	ND	>50000	inactive	inactive	98.7 ± 1.0	3805 ± 2944	12	>5	2.0	>26
16	44.1 ± 6.3	1455 ± 572	77.7 ± 3.7	7067 ± 2290	41.4 ± 5.83	1601 ± 867	ND	>50000	inactive	inactive	97.9 ± 1.2	1086 ± 597	ND	>1	7.0	>46
17	inactive	inactive	94.6 ± 5.4	11847 ± 2000	inactive	inactive	ND	>50000	inactive	inactive	97.1 ± 0.8	430 ± 195	ND	ND	28	>116
12	inactive	inactive	100 ± 0	>100 000	inactive	inactive	inactive	inactive	inactive	inactive	100 ± 0	7777 ± 2166	ND	ND	>13	ND
13	inactive	inactive	100 ± 0	>100 000	inactive	inactive	inactive	inactive	16.4 ± 3.9	9212 ± 6238	96.7 ± 2.7	4252 ± 1077	ND	ND	>24	ND
20	inactive	inactive	100 ± 0	67613 ± 17748	inactive	inactive	inactive	inactive	inactive	inactive	91.7 ± 6.8	2622 ± 678	ND	ND	26	ND
21	18.7 ± 0.4	3887 ± 1878	100 ± 0	88447 ± 11553	44.7 ± 5.9	2755 ± 472	100 ± 0	88 205 ± 9631	26.2 ± 5.1	133 ± 60.3	59.7 ± 4.6	370 ± 105	29	21	239	238
22	inactive	inactive	100 ± 0	70703 ± 29297	inactive	inactive	inactive	inactive	inactive	inactive	80.4 ± 6.0	10449 ± 2482	ND	ND	6.8	ND
25	21.3 ± 6.8	48503 ± 27571	inactive	inactive	18.6 ± 2.7	2802 ± 2507	inactive	inactive	19.0 ± 2.7	2657 ± 121	100 ± 0	25780 ± 9773	18	1.0	ND	ND
26	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	93.1 ± 7	23450 ± 11959	ND	ND	ND	ND
27	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	ND	ND	ND	ND
28	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	inactive	100 ± 0	>100000	ND	ND	ND	ND

Values determined by nonlinear regression of individual experiments run in triplicate as detailed in materials and methods under β-arrestin assays. All EC50, IC50, and Emax values are presented as means ± SEM; n = 3–4. ND indicates not determined due to an incomplete curve. Inactive indicates no measurable activity in indicated assay.

Replacing the piperidinyl ring of 1 with a piperazine (Figure , class 1)—typified by 10 and 21—resulted in similar binding and agonist efficacy profiles at D4R, improved subtype selectivity (Tables and 2), and a gain in efficacy at both D2R and D3R (Figure and Tables and 3). Replacing the pyridinyl ring of 1 with a phenyl or napthyl moiety (Figure , class 2)—typified by 6 and 13—resulted in improved subtype selectivity, and importantly a diminished-efficacy partial agonist profile at D4R. These compounds showed no measurable agonist efficacy at either D2R or D3Rs (Figure ). A para-substitution on the pyridinyl ring of 1 (Figure , class 3)—typified by 12 and 9—resulted in compounds that lost all agonist efficacy but retained high-affinity binding at D4R, with very minimal binding at D2R or D3R. The compounds showed potent antagonism of the D4R response with minimal low potency D2R blockade and no measurable affinity or efficacy at D3R. Therefore, this class of compounds represents highly selective D4R antagonists with no measurable agonist efficacy on any D2-like receptor (Figure , Tables –3).

Figure 4

Compounds 13 (yellow) and 6 (blue) show diminished agonist activity at the D4R compared to parent compound 1 (black). D4R-expressing stable cells lines were plated and compounds were assayed for agonist (A) and antagonist (B) activity on β-arrestin recruitment. Similarly, D4R-mediated inhibition of cAMP accumulation was also examined in both agonist (C), and antagonist (D) modes, as indicated. Assays were conducted as described in the Experimental Methods; briefly, agonist assays were conducted by incubating the cells with the indicated concentration of test compound and measuring luminescence. Antagonist assays were conducted by incubating the compound with an EC80 concentration of dopamine (1 μM for β-arrestin and 10 nM in cAMP) and the indicated concentration of test compound. For cAMP assays, cells were first stimulated with 10 μM forskolin. Agonist mode assays are expressed as a percentage of the maximum dopamine response, whereas antagonist mode assays are expressed as a percentage of dopamine’s EC80 response. Emax and EC50 values are shown in Tables and 3. Dopamine and sulpiride were run during each assay as positive controls for a full agonist and full antagonist, respectively (data not shown). Data were fit using nonlinear regression of individual experiments performed in triplicate and are shown as means ± SEM; n = 3.

Figure 5

Compounds 12 (green) and 9 (purple) are full antagonists at the D4R. D4R-expressing stable cells lines were plated and compounds were assayed for agonist (A) and antagonist (B) activity on β-arrestin recruitment. Similarly, D4R-mediated inhibition of cAMP accumulation was also examined in both agonist (C), and antagonist (D) modes, as indicated. Assays were conducted as described in the Experimental Methods; briefly, agonist assays were conducted by incubating the cells with the indicated concentration of test compound and measuring luminescence. Antagonist assays were conducted by incubating the compound with an EC80 concentration of dopamine (1 μM for β-arrestin and 10 nM in cAMP) and the indicated concentration of test compound. For cAMP assays, cells were first stimulated with 10 μM forskolin. Assays were conducted as described in the Experimental Methods. Agonist mode assays are expressed as a percentage of the maximum dopamine response, whereas antagonist mode assays are expressed as a percentage of dopamine’s EC80 (1 μM in β-arrestin and 10 nM in cAMP) response. Emax and EC50 values are shown in Tables and 3. Dopamine and sulpiride were run during each assay as positive controls for a full agonist and full antagonist respectively (data not shown). Data were fit using nonlinear regression of individual experiments performed in triplicate and are shown as means ± SEM; n = 3.

Individual compounds within classes 1–3 resulted in modest changes to overall efficacy and potency as overviewed in Tables –3. For this reason, we chose to focus on typified examples of a range of agonist efficacy (higher, medium, and none) at the D4R. Using these classes, we performed MD simulations to identify interaction sites on the receptor that may play a pivotal role in engendering agonist selectivity and efficacy.

MD Studies

To gain insights on probable ligand interactions at D4R, a set of seven ligands from the parent compound and the three class of modifications (i.e., 1, 6, 9, 10, 12, 13, and 21) were docked to the crystal structures of D2R,[26] D3R,[25] and D4R.[24] Each receptor–ligand combination was subjected to 100 ns MD simulations, followed by the simulation interaction diagram (SID) and clustering analysis as described in the Experimental Methods section. The results are included in the Supporting Information (Tables S2–S4 and Figures S2–S39). Comparisons of structural and dynamic properties of each ligand–receptor system, with reference to the parent compound 1 for each receptor system, are listed in Table S1. Although the same class modifications caused similar changes in the majority of the analyzed properties, some subtle differences are also identified. A representative ligand–receptor system from each class modification is presented here.

In order to explore class 1 modifications that showed a gain of efficacy at D2R and D3R with minimal changes in D4R binding or efficacy, 10 was selected to be presented here along with parent compound 1. The comparative ligand binding at of 10 D2R (Figure ) and D3R (Figure ) revealed that the modest ligand change—the substitution of a piperazine for a piperidine—induced a dramatic shift in the binding orientation at D2R and D3R: compared to the parent compound 1, 10 took on a rotated orientation in both receptors, in which the arylamide portion of the 10 occupies a region of the binding pocket that accommodates the 2-(piperidin-4-yl)pyridinyl portion of 1. This pose allows 10 to better engage with the conserved transmembrane (TM) 3 aspartate residue (D3.32) located within the orthosteric binding pocket of biogenic amine receptors like dopaminergic receptors.[32] Additionally, there was new engagement with conserved V2.61, and additional TM5 and TM6 helix shifts in both receptors. In contrast, the binding orientation of 10 at D4R is similar to that of 1 (Figure S7), although a shift in the orientation of the pyridinylpiperidine ring system deeper into the receptor was observed. 21 docked similarly to 10 at D2R and at D4R (i.e., rotated 180° in comparison to 1), but differed at D3R in which the pose was similar to that of 1, possibly indicating a different activation mechanism for D3R by this compound.

Figure 6

1 and 10 docked at D2R. (A–D) Comparative alignment of 1 (red ligand, yellow TM domains) and 10 (blue ligand, purple TM domains) following MD simulations of the D2R (PDB: 6CM4(26)). (E–H) Analysis of ligand interactions with specific side chains of 1 (E,G) and 10 (F,H). Although the structural difference between 1 and 10 is only a piperidine vs a piperazine ring, this drives a dramatic shift in ligand orientation in which 10 is “flipped” and rotated by 180° about its longitudinal axis, with its pyridine ring deepest in the binding pocket. This allows the basic nitrogen of the neighboring piperazine ring to engage the conserved aspartate in TM3, a common feature of dopaminergic agonists.

Figure 7

1 and 10 docked at D3R. (A–D) Comparative alignment of 1 (red ligand, yellow TM domains) and 10 (blue ligand, purple TM domains) following MD simulations of the D3R (PDB: 3PBL(25)). (E–H) Analysis of ligand interactions with specific side chains of 1 (E,G) and 10 (F,H). As seen in the D2R model, 10 is also “flipped” and rotated by 180° about its longitudinal axis in the binding pocket at D3R compared to 1. This allows for a different set of hydrophobic interactions and the engagement of the basic nitrogen of the piperazine ring to with the conserved aspartate in TM3.

13 (Figure ), representing class 2 modifications that showed a partial loss of efficacy at D4R, and 9 (Figure ), representing class 3 modifications that showed a complete loss of efficacy at D4R, are shown in models of D4R alongside the parent compound 1. 13 uniquely engaged with S2.64, E2.65, and T7.39, and induced conformational shifts in several TM domains and intra/extracellular loops. 9 showed grater engagement with ECL2 and uniquely interacted with C3.25 and W6.48. Whereas 13 adopted a pose similar to 1, 9 adopted a rotated orientation in which the arylamide portion of the 9 occupies a region of the binding pocket that accommodates the 2-(piperidin-4-yl)pyridinyl portion of 1 (Table S1). The results seen in these two comparisons are consistent with previous observations in which regiosubstitutions on an aryl ring of a terminal arylpiperazine can modulate efficacy at D4R.[33,34] In particular, the inclusion of a para substitution on the terminal arylpiperazine has reliably produced D4R antagonists for a wide variety of molecules with diverse substituents on the secondary pharmacophore.

Figure 8

1 and 13 docked at D4R. (A–D) Comparative alignment of 1 (red ligand, yellow TM domains) and 13 (blue ligand, purple TM domains) following MD simulations of the D4R (PDB: 5WIU(24)). (E–H) Analysis of ligand interactions with specific side chains of 1 (E,G) and 13 (F,H). The bulky napthyl ring of 13 shifts the overall fit within the extended binding pocket, partially disrupting the engagement of the basic nitrogen of the piperazine ring to with the conserved aspartate in TM3.

Figure 9

1 and 9 docked at D4R. (A–D) Comparative alignment of 1 (red ligand, yellow TM domains) and 9 (blue ligand, purple TM domains) following MD simulations of the D4R (PDB: 5WIU(24)). (E–H) Analysis of ligand interactions with specific side chains of 1 (E,G) and 9 (F,H). The inclusion of a single para substitution on the pyridine ring of 9 induces a “flipped” orientation of the ligand, in which the binding pose is rotated by 180° about its longitudinal axis, with its pyridine ring deepest in the binding pocket driving the arylamide into a deeper binding position.

Conclusions

Evidence from human genetic studies and animal models suggest that D4R signaling may mediate behavioral traits including impulsivity,[35] novelty seeking,[35−38] fear and anxiety,[39,40] and sensitivity to drugs of abuse.[40−43] While modulation of postsynaptic D4R expression in the PFC is typically hypothesized to mediate the reported in vivo effects of D4R agonists and antagonists, evidence suggests important roles of D4R expression in the nucleus accumbens shell[44] and within the lateral habenula,[45] in which the receptor may be preferentially activated by norepinephrine rather than dopamine.[46,47] Furthermore, little is known about the physiological relevance of independent D4R-mediated signaling pathways (e.g., cAMP and β-arrestin) in the manifestation of behavioral outputs. A recent report identified a D4R-selective compound containing an unsubstituted phenylpiperazine that potently and partially activated Gαi but inhibited β-arrestin2 recruitment and identified likely ligand–residue interactions that affect receptor activation states.[48] There is much left to be determined about the physiological role of D4R signaling in modulating attention and cognitive processes, and new selective agonists and antagonists of these receptors will be valuable tools for deduction of signaling importance by these receptors.

New highly selective D4R partial agonists and antagonists will be useful to better characterize the role of D4R signaling in vivo. While we have primarily focused on selectivity against the very closely related D2R and D3R, it will be important to establish global selectivity of these compounds for in vivo experimentation through a broader screening of biogenic amine receptors. To this end, we investigated a subset of these compounds on the related D1-like DARs. None of the tested compounds showed any measurable agonist or antagonist effect at either the D1R or D5R (Figure S1). Comprehensive binding and functional studies, in concert with detailed molecular modeling analyses using newly published crystal structures, provides a platform for developing high-affinity and highly subtype selective ligands of varying efficacies. This study aimed to identify key molecular interactions that dictate D4R potency, efficacy, and subtype selectivity.

The parent compound, 1, was confirmed to be a high-affinity (low nM Ki value) partial agonist (Emax = 23–62%) at D4R. Illustrative of broader trends from our library, the compound Ki values for all D2-like receptors determined using [3H]-(R)-(+)-7-OH-DPAT competition assays tended to be lower (more potent) than those obtained using [3H]N-methylspiperone. As expected, the divergence between these values increased with agonist efficacy, consistent with previous experience regarding agonist versus antagonist radioligands.[30]

Key modifications to the 1 pharmacophore provided modest gains to D4R affinity, but dramatic gains in selectivity over D2R and D3R, likely because of a substantial decrease in D2R and D3R engagement by the analogs. Interestingly, the manner of these substitutions produced three classes of lead compound: (1) those with binding and efficacy profiles similar to 1 at D4R but gains in efficacy at D2R and D3R; (2) those with improved D4R binding and subtype selectivity with lower partial agonist efficacy; and (3) those with improved D4R binding and subtype selectivity with full antagonist characteristics. MD simulations suggest that the gain in D2R and D3R efficacy seen in compounds like 10 could be partially due to achieving a rotated ligand pose that more fully engages the conserved TM3 aspartate. Similarly, the complete shift to antagonism at D4R seen in compounds like 9 could be partially due to inducing an alternate binding pose that either no longer allows full engagement of the orthosteric binding site and occupation of an alternative secondary binding pocket or a ligand-dependent alteration of the receptor energy landscape leading to the stabilization of a different receptor conformation.

These molecular models provide testable predictions relative to the unique interaction sites of these diverse compounds within the D4R. These interactions likely underlie agonist efficacy of a given compound. Interestingly, compounds can be “rank ordered” by levels of agonist efficacy starting with 1 having the highest D4R activation, followed by class 1 compounds (which show similar agonist efficacy) and then class 2 compounds (which show less agonist efficacy) and class 3 compounds that lack any agonist efficacy. As expected, these compounds align the opposite way for antagonist efficacy, wherein a lower agonist efficacy correlates with a higher antagonist efficacy. Examined this way, one can see that it may be possible to “dial-in” or “dial-out” levels of D4R stimulation via adjusting compound structure, and therefore interaction sites on the receptor, leading to divergent levels of partial agonism. Future studies will involve further SAR and receptor mutagenesis studies to verify these models. We are optimistic that some of the analogues may be developed into useful in vivo research tools and plan to examine absorption, distribution, metabolism, and excretion characteristics of selected analogues. It is interesting to speculate that a collection of partial agonists with varying efficacies may allow for the fine-tuning of D4R activation, potentially leading to a fuller understanding of functional consequences of varying signaling levels for D4R-targeted therapeutics for neuropsychiatric disorders.

Experimental Methods

Synthesis

Reaction conditions and yields were not optimized. Anhydrous solvents were purchased from Aldrich and were used without further purification. All other chemicals and reagents were purchased from Sigma-Aldrich Co. LLC, Combi-Blocks, TCI America, OChem Incorporation, Acros Organics, and Alfa Aesar. All amine final products were converted into either the oxalate or hydrochloride salt. Spectroscopic data and yields refer to the free base form of compounds. Flash chromatography was performed using silica gel (EMD Chemicals, Inc.; 230–400 mesh, 60 Å) by using a Teledyne ISCO CombiFlash RF system. 1H NMR spectra were acquired using a Varian Mercury Plus 400 spectrometer at 400 MHz. Chemical shifts are reported in parts-per-million and referenced according to deuterated solvent for 1H spectra (CDCl3, 7.26, CD3OD, 3.31 or D2O, 4.79). Combustion analysis was performed by Atlantic Microlab, Inc., (Norcross, GA), and the results agree within ±0.4% of calculated values (Table S5). Melting point determination was conducted using a Stanford Research Systems OptiMelt automated melting point apparatus and are uncorrected. On the basis of NMR and combustion data, all final compounds are >95% pure. All compounds within this series are covered under an existing patent,[49] but only 1,[21,22]6,[22] and 10(16,21) have been previously described in the peer-reviewed literature.

General Method A

2-Chloro-N-(m-tolyl)acetamide (5)[21,27,28]

2-Chloroacetyl chloride (1.16 g, 10.3 mmol) was added to a solution of m-toluidine (1.00 mL, 9.33 mmol) in ethyl acetate (30 mL) and triethylamine (1.43 mL, 10.3 mmol) at 0 °C. The reaction mixture was allowed to warm to room temperature and stirred overnight under a N2 atmosphere. After the reaction was complete, the solvent was removed in vacuo. The crude mixture was diluted with water (100 mL) and EtOAc (100 mL) and then extracted with EtOAc (3 × 100 mL) and washed with brine (100 mL). The combined organic layer was dried over MgSO4, filtered, and concentrated. The product was purified by column chromatography (10–90% EtOAc/hexanes gradient) to give 5 (1.30 g, 76% yield) as an off-white solid. 1H NMR (CDCl3): δ 8.19 (s, 1H), 7.38–7.33 (m, 1H), 7.26–7.22 (m, 1H), 6.99 (d, J = 7.6 Hz, 1H), 4.18 (s, 2H), 2.36 (s, 3H).

General Method B

2-(4-Phenylpiperidin-1-yl)-N-(m-tolyl)acetamide (6; CAB03-015)[21,22,27,28]

K2CO3 (2.57 g, 18.6 mmol) and NaI (50 mg) were added to a solution of 2-chloro-N-(m-tolyl)acetamide (570 mg, 3.10 mmol) and commercially available 4-phenylpiperidine (500 mg, 3.10 mmol) in an anhydrous acetonitrile (12 mL) solution. The reaction mixture was stirred at reflux (80 °C) for 20 h under a N2 atmosphere. The reaction mixture was cooled to room temperature, and the solvent was removed in vacuo. The residue was diluted with water (100 mL) and EtOAc (100 mL) and then extracted with EtOAc (3 × 100 mL) and washed with brine (100 mL). The combined organic layer was dried over MgSO4, filtered, and concentrated. The crude product was purified by column chromatography (10–90% EtOAc/hexanes gradient) to give pure product 6 (710 mg, 74% yield) as an off-white solid. mp 70–71 °C; 1H NMR (CDCl3): δ 9.15 (s, 1H), 7.42–7.40 (m, 2H), 7.38–7.31 (m, 2H), 7.24–7.20 (m, 4H), 6.93 (d, J = 7.2 Hz, 1H), 3.15 (s, 2H), 3.10–3.02 (m, 2H), 2.59–2.52 (m, 1H), 2.43–2.38 (m, 2H), 2.36 (s, 3H), 1.94–1.90 (m, 4H). Anal. (C20H24N2O·HCl·1/4H2O) C, H, N.

N-(m-Tolyl)-2-(4-(p-tolyl)piperidin-1-yl)acetamide (7; CAB02-007HP)

Compound 7 was synthesized as described for 6 using K2CO3 (1.18 g, 8.56 mmol), NaI (40.0 mg) 4-(p-tolyl)piperidine (250 mg, 1.43 mmol), and 2-chloro-N-(m-tolyl)acetamide (262 mg, 1.43 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (15–85% EtOAc/hexanes gradient) to give pure product 7 (182 mg, 40% yield) as a white solid. mp 63–65 °C; 1H NMR (400 MHz, CDCl3): δ 9.23 (s, 1H), 7.47 (d, J = 15.7 Hz, 2H), 7.31 (dd, J = 14.4, 5.5 Hz, 2H), 7.23 (d, J = 2.9 Hz, 3H), 7.00 (d, J = 7.5 Hz, 1H), 3.22 (s, 2H), 3.10 (d, J = 11.4 Hz, 2H), 2.58 (d, J = 12.5 Hz, 1H), 2.49–2.46 (m, 2H), 2.43 (s, 3H), 2.41 (s, 3H), 1.91 (dt, J = 37.8, 13.1 Hz, 4H). Anal. (C21H26N2O·1/4H2O·2/5C3H8O) C, H, N.

2-(4-(4-Chlorophenyl)piperidin-1-yl)-N-(m-tolyl)acetamide (8; CAB02-009HP)

Compound 8 was synthesized as described for 6 using K2CO3 (715 mg, 5.17 mmol), NaI (40.0 mg) 4-(4-chlorophenyl)piperidine (200 mg, 0.86 mmol), and 2-chloro-N-(m-tolyl)acetamide (158 mg, 0.86 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (40–60% EtOAc/hexanes gradient) to give pure product 8 (120 mg, 41% yield) as a light brownish solid. mp 114–116 °C; 1H NMR (400 MHz, CDCl3): δ 9.08 (s, 1H), 7.42–7.33 (m, 2H), 7.28 (d, J = 8.5 Hz, 1H), 7.24 (d, J = 2.4 Hz, 2H), 7.16 (d, J = 8.1 Hz, 2H), 6.95–6.88 (m, 1H), 3.13 (s, 2H), 3.01 (t, J = 7.1 Hz, 2H), 2.50 (d, J = 12.4 Hz, 1H), 2.35 (s, 2H), 2.34 (s, 3H), 1.88 (d, J = 13.1 Hz, 2H), 1.76 (q, J = 12.7 Hz, 2H). Anal. (C20H23ClN2O) C, H, N.

2-(4-(5-Methylpyridin-2-yl)piperidin-1-yl)-N-(m-tolyl)acetamide (9; CAB02-005HP)

Compound 9 was synthesized as described for 6 using K2CO3 (1.64 g, 11.9 mmol), NaI (40.0 mg) 5-methyl-2-(piperidin-4-yl)pyridine (350 mg, 1.98 mmol), and 2-chloro-N-(m-tolyl)acetamide (363 mg, 1.98 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (50–50% EtOAc/hexanes gradient) to give pure product 9 (420 mg, 66% yield) as a white solid. mp 126–128 °C; 1H NMR (400 MHz, CDCl3): δ 9.09 (s, 1H), 8.03 (s, 1H), 7.43–7.30 (m, 3H), 7.20 (d, J = 8.2 Hz, 1H), 6.93 (s, 1H), 6.61 (d, J = 8.5 Hz, 1H), 3.62–3.49 (m, 3H), 3.18 (br s, 2H), 2.74 (d, J = 5.8 Hz, 4H), 2.34 (s, 3H), 2.20 (s, 3H), 1.58–1.52 (m, 2H).

2-(4-(Pyridin-2-yl)piperazin-1-yl)-N-(m-tolyl)acetamide (10; CAB02-140)[22,27]

Compound 10 was synthesized as described for 6 using K2CO3 (1.13 g, 6.78 mmol), NaI (40.0 mg) 1-(pyridin-2-yl)piperazine (222 mg, 1.36 mmol), and 2-chloro-N-(m-tolyl)acetamide (250 mg, 1.36 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 10 (330 mg, 78% yield) as a white solid. mp 127–129 °C; 1H NMR (400 MHz, CDCl3): δ 9.09 (s, 1H), 8.21 (d, J = 4.6 Hz, 1H), 7.56–7.46 (m, 1H), 7.39 (d, J = 10.1 Hz, 2H), 7.24–7.19 (m, 1H), 6.94 (d, J = 7.4 Hz, 1H), 6.68 (q, J = 5.5, 4.4 Hz, 2H), 3.62 (t, J = 5.0 Hz, 4H), 3.19 (d, J = 3.2 Hz, 2H), 2.74 (d, J = 5.4 Hz, 4H), 2.35 (d, J = 3.2 Hz, 3H), 2.17 (br s, 4H). Anal. (C18H22N4O·2HCl·H2O) C, H, N.

2-(4-(Pyrimidin-2-yl)piperazin-1-yl)-N-(m-tolyl)acetamide (11; CAB02-110)

Compound 11 was synthesized as described for 6 using K2CO3 (1.13 g, 1.16 mmol), 2-(piperazin-1-yl)pyrimidine (223 mg, 1.36 mmol), and 2-chloro-N-(m-tolyl)acetamide (250 mg, 1.36 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (10–90% EtOAc/hexanes gradient) to give pure product 11 (310 mg, 73% yield) as a cream solid. mp 92–94 °C; 1H NMR (400 MHz, CDCl3): δ 9.16–9.01 (m, 1H), 8.37–8.27 (m, 2H), 7.39 (d, J = 14.2 Hz, 2H), 7.22 (d, J = 8.1 Hz, 1H), 6.94 (d, J = 8.4 Hz, 1H), 6.53 (d, J = 4.6 Hz, 1H), 3.91 (s, 4H), 3.18 (q, J = 2.1, 1.6 Hz, 2H), 2.69 (d, J = 5.6 Hz, 4H), 2.36 (s, 3H). Anal. (C17H21N5O·2HCl·1.75H2O) C, H, N.

2-(4-(5-Chloropyridin-2-yl)piperazin-1-yl)-N-(m-tolyl)acetamide (12; CAB02-003HP)

Compound 12 was synthesized as described for 6 using K2CO3 (1.78 g, 12.86 mmol), NaI (40.0 mg), 1-(5-chloropyridin-2-yl)piperazine (422 mg, 2.14 mmol), and 2-chloro-N-(m-tolyl)acetamide (400 mg, 2.14 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 12 (490 mg, 65% yield) as a white solid. mp 107–109 °C; 1H NMR (400 MHz, CDCl3): δ 9.02 (s, 1H), 8.11 (d, J = 3.3 Hz, 1H), 7.47–7.40 (m, 1H), 7.36 (d, J = 11.8 Hz, 2H), 7.19 (d, J = 8.5 Hz, 1H), 6.95–6.88 (m, 1H), 6.59 (d, J = 8.9 Hz, 1H), 3.58 (t, J = 5.2 Hz, 4H), 3.17 (d, J = 2.8 Hz, 2H), 2.71 (t, J = 5.0 Hz, 4H), 2.33 (s, 3H). Anal. (C18H21ClN4O·2HCl·1/2H2O) C, H, N.

2-(4-(Naphthalen-1-yl)piperazin-1-yl)-N-(m-tolyl)acetamide (13; CAB02-011HP)

Compound 13 was synthesized as described for 6 using K2CO3 (2.62 g, 18.9 mmol), NaI (50.0 mg), 1-(naphthalen-1-yl)piperazine[29] (670 mg, 3.16 mmol), and 2-chloro-N-(m-tolyl)acetamide (580 mg, 3.16 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (40–60% EtOAc/hexanes gradient) to give pure product 13 (546 mg, 48% yield) as a brown oil; 1H NMR (400 MHz, CDCl3): δ 9.17 (s, 1H), 8.20 (d, J = 7.5 Hz, 1H), 7.85 (s, 1H), 7.60 (d, J = 8.2 Hz, 1H), 7.55–7.47 (m, 2H), 7.43 (d, J = 11.4 Hz, 3H), 7.24 (d, J = 8.4 Hz, 1H), 7.15 (d, J = 7.3 Hz, 1H), 6.96 (d, J = 7.4 Hz, 1H), 3.34–3.27 (m, 2H), 3.23 (br s, 4H), 2.95 (br s, 4H), 2.38 (s, 3H). Anal. (C23H25N3O·2HCl) C, H, N.

3-Chloro-N-(m-tolyl)propanamide (14)

Compound 14 was synthesized as described for 5 by adding 3-chloropropanoyl chloride (1.30 g, 10.3 mmol) to a solution of m-toluidine (1 mL, 9.33 mmol) in ethyl acetate (30 mL) and triethylamine (1.44 mL). The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give compound 14 (1.43 g, 78% yield) as a white solid. 1H NMR (CDCl3): δ 7.44 (s, 1H), 7.36–7.22 (m, 2H), 6.99 (d, J = 7.4 Hz, 1H), 3.96–3.92 (m, 2H), 2.85 (t, J = 6.3 Hz, 2H), 2.39 (s, 3H).

3-(4-(Pyridin-2-yl)piperidin-1-yl)-N-(m-tolyl)propanamide (15; CAB02-120)

Compound 15 was synthesized as described for 6 using K2CO3 (1.26 g, 7.56 mmol), NaI (50.0 mg), 2-(piperidin-4-yl)pyridine (246 mg, 1.52 mmol), and 3-chloro-N-(m-tolyl)propanamide (300 mg, 1.52 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (50–50% EtOAc/hexanes gradient) to give pure product 15 (325 mg, 66% yield) as a light brown oil; 1H NMR (400 MHz, CDCl3): δ 11.08 (s, 1H), 8.55 (d, J = 4.6 Hz, 1H), 7.69–7.61 (m, 1H), 7.45 (s, 1H), 7.29 (d, J = 7.5 Hz, 1H), 7.23–7.12 (m, 3H), 6.88 (d, J = 7.5 Hz, 1H), 3.21 (d, J = 11.2 Hz, 2H), 2.87–2.70 (m, 3H), 2.53 (t, J = 5.7 Hz, 2H), 2.34 (s, 3H), 2.24 (t, J = 11.6 Hz, 2H), 2.09 (d, J = 13.2 Hz, 2H), 2.02–1.87 (m, 2H). Anal. (C20H25N3O·2HCl·3.5H2O) C, H, N.

3-(4-(Pyridin-2-yl)piperazin-1-yl)-N-(m-tolyl)propanamide (16; CAB02-142)

Compound 16 was synthesized as described for 6 using K2CO3 (1.26 g, 9.12 mmol), 1-(pyridin-2-yl)piperazine (248 mg, 1.52 mmol), and 3-chloro-N-(m-tolyl)propanamide (300 mg, 1.52 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (50–50% EtOAc/hexanes gradient) to give pure product 16 (350 mg, 71% yield) as a white solid. mp 99–101 °C; 1H NMR (400 MHz CDCl3): δ 10.74 (s, 1H), 8.21 (br s, 1H), 7.51 (d, J = 8.0 Hz, 1H), 7.40 (s, 1H), 7.24 (d, J = 5.7 Hz, 1H), 7.15 (td, J = 8.2, 3.3 Hz, 1H), 6.86 (d, J = 7.3 Hz, 1H), 6.68 (q, J = 7.4, 5.6 Hz, 2H), 3.64 (d, J = 6.4 Hz, 4H), 2.81–2.65 (m, 6H), 2.55 (q, J = 4.7 Hz, 2H), 2.30 (s, 3H). Anal. (C19H24N4O·3HCl) C, H, N.

3-(4-(Pyrimidin-2-yl)piperazin-1-yl)-N-(m-tolyl)propanamide (17; RNB01-007)

Compound 17 was synthesized as described for 6 using K2CO3 (1.26 g, 9.12 mmol), 2-(piperazin-1-yl)pyrimidine (249 mg, 1.52 mmol), and 3-chloro-N-(m-tolyl)propanamide (350 mg, 1.52 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (50–50% EtOAc/hexanes gradient) to give pure product 17 (321 mg, 56% yield) as a cream solid. mp 85–86 °C; 1H NMR (400 MHz, CDCl3): δ 10.76 (s, 1H), 8.33 (dt, J = 5.1, 2.4 Hz, 2H), 7.40 (s, 1H), 7.32–7.23 (m, 1H), 7.21–7.13 (m, 1H), 6.88 (d, J = 7.5 Hz, 1H), 6.54 (dp, J = 5.0, 2.6, 2.1 Hz, 1H), 3.94 (t, J = 4.8 Hz, 4H), 2.77 (t, J = 5.7 Hz, 2H), 2.67 (t, J = 4.6 Hz, 4H), 2.57 (q, J = 5.7, 5.0 Hz, 2H), 2.32 (s, 3H). Anal. (C18H23N5O·HCl) C, H, N.

2-Chloro-N-(3-ethylphenyl)acetamide (19)

Compound 19 was synthesized as described for 5 by adding 2-chloroacetyl chloride (3.62 mL, 45.4 mmol) to a solution of 3-ethylaniline (5.13 mL, 41.3 mmol) in ethyl acetate (30 mL) and triethylamine (1.44 mL). The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give compound 19 (8.00 g, 98% yield) as a cream solid. 1H NMR (400 MHz, CDCl3): δ 8.23 (s, 1H), 7.47–7.38 (m, 2H), 7.32–7.27 (m, 1H), 7.04 (d, J = 7.5 Hz, 1H), 4.20 (s, 2H), 2.75–2.58 (m, 2H), 1.28–1.22 (m, 3H).

N-(3-Ethylphenyl)-2-(4-(pyridin-2-yl)piperidin-1-yl)acetamide (20; CAB02-021HP)

Compound 20 was synthesized as described for 6 using K2CO3 (1.54 g, 11.17 mmol), 2-(piperidin-4-yl)pyridine (300 mg, 1.86 mmol), and 2-chloro-N-(3-ethylphenyl)acetamide (368 mg, 1.86 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (40–60% EtOAc/hexanes gradient) to give pure product 20 (435 mg, 73% yield) as a white solid. mp 56–58 °C; 1H NMR (400 MHz, CDCl3): δ 9.16 (s, 1H), 7.48–7.38 (m, 2H), 7.33 (d, J = 7.4 Hz, 2H), 7.26 (br s, 4H), 6.97 (d, J = 7.5 Hz, 1H), 3.16 (d, J = 3.3 Hz, 2H), 3.05 (d, J = 11.5 Hz, 2H), 2.66 (d, J = 7.9 Hz, 2H), 2.60–2.54 (m, 1H), 2.40 (t, J = 12.0 Hz, 2H), 1.87 (dt, J = 38.0, 13.2 Hz, 4H), 1.25 (t, J = 7.3, 5.9 Hz, 3H). Anal. (C20H25N3O·C2H2O4) C, H, N.

N-(3-Ethylphenyl)-2-(4-(pyridin-2-yl)piperazin-1-yl)acetamide (21; CAB02-017HP)

Compound 21 was synthesized as described for 6 using K2CO3 (2.74 g, 19.8 mmol), 1-(pyridin-2-yl)piperazine (539 mg, 3.30 mmol), and 2-chloro-N-(3-ethylphenyl)acetamide (650 mg, 3.30 mmol) in an anhydrous acetonitrile (8 mL) solution. The crude product was purified by column chromatography (50–50% EtOAc/hexanes gradient) to give pure product 21 (792 mg, 74% yield) as a white solid. mp 109–111 °C; 1H NMR (400 MHz, CDCl3): δ 9.07 (s, 1H), 8.19 (s, 1H), 7.54–7.44 (m, 1H), 7.38 (d, J = 8.0 Hz, 2H), 7.27–7.20 (m, 1H), 6.95 (d, J = 7.6 Hz, 1H), 6.66 (d, J = 8.6 Hz, 2H), 3.60 (d, J = 6.0 Hz, 4H), 3.17 (t, J = 2.3 Hz, 2H), 2.73 (d, J = 5.0 Hz, 4H), 2.61 (t, J = 8.0 Hz, 2H), 1.26–1.15 (m, 3H). Anal. (C19H24N4O·2HCl·3/4H2O) C, H, N.

2-(4-(5-Chloropyridin-2-yl)piperazin-1-yl)-N-(3-ethylphenyl)acetamide (22; CAB02-019HP)

Compound 22 was synthesized as described for 6 using K2CO3 (1.26 g, 9.12 mmol), 1-(5-chloropyridin-2-yl)piperazine (300 mg, 1.52 mmol), and 2-chloro-N-(3-ethylphenyl)acetamide (300 mg, 1.52 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 22 (480 mg, 88% yield) as a white solid. mp 101–103 °C; 1H NMR (400 MHz, CDCl3): δ 9.03 (s, 1H), 8.11 (s, 1H), 7.46–7.34 (m, 3H), 7.25–7.23 (m, 1H), 6.95 (d, J = 7.5 Hz, 1H), 6.59 (d, J = 8.9 Hz, 1H), 3.57 (t, J = 4.9 Hz, 4H), 3.17 (d, J = 2.7 Hz, 2H), 2.71 (t, J = 4.7 Hz, 4H), 2.61 (t, J = 8.1 Hz, 2H), 1.32–1.13 (m, 3H). Anal. (C19H23ClN4O·2HCl·3/4H2O) C, H, N.

2-Chloro-N-(pyridin-3-yl)acetamide (24a)

Compound 24a was synthesized as described for 5 by adding 2-chloroacetyl chloride (0.71 mL, 17.5 mmol) to a solution of pyridin-3-amine (1.50 g, 16.0 mmol) in ethyl acetate (25 mL) and triethylamine (0.4 mL). The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give 24a (1.17 g, 43% yield) as a white solid. 1H NMR (400 MHz, D2O): δ 9.37–9.08 (m, 1H), 8.60–8.32 (m, 2H), 8.06–7.87 (m, 1H), 5.67 (s, 2H).

2-Chloro-N-(pyrimidin-5-yl)acetamide (24b)

Compound 24b was synthesized as described for 5 by adding 2-chloroacetyl chloride (0.46 mL, 5.78 mmol) to a solution of pyrimidin-5-amine (500 mg, 5.26 mmol) in ethyl acetate (10 mL) and triethylamine (0.4 mL). The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give 24b (510 mg, 57% yield) as a brown solid. 1H NMR (400 MHz, D2O): δ 9.17 (s, 1H), 8.54 (s, 2H), 8.01 (s, 1H), 5.66 (s, 2H).

2-(4-(Pyridin-2-yl)piperazin-1-yl)-N-(pyridin-3-yl)acetamide (25; CAB02-033HP)

Compound 25 was synthesized as described for 6 using K2CO3 (1.27 g, 9.17 mmol), 1-(pyridin-2-yl)piperazine (249 mg, 1.52 mmol), and compound 24a 2-chloro-N-(pyridin-3-yl)acetamide (260 mg, 1.52 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 25 (287 mg, 63% yield) as a white solid. mp 173–174 °C; 1H NMR (400 MHz CDCl3): δ 9.23 (s, 1H), 8.59 (s, 1H), 8.37 (s, 1H), 8.23 (d, J = 12.8 Hz, 2H), 7.51 (s, 1H), 6.68 (d, J = 8.1 Hz, 2H), 3.63 (br s, 4H), 3.23 (d, J = 2.9 Hz, 2H), 2.76 (br s, 4H). Anal. (C16H19N5O·4HCl·1.1H2O·1/4C3H8O) C, H, N.

2-(4-(5-Chloropyridin-2-yl)piperazin-1-yl)-N-(pyridin-3-yl)acetamide (26; CAB02-035HP)

Compound 26 was synthesized as described for 6 using K2CO3 (2.63 g, 19.0 mmol), 1-(5-chloropyridin-2-yl)piperazine (627 mg, 3.17 mmol), and compound 24a 2-chloro-N-(pyridin-3-yl)acetamide (540 mg, 3.17 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 26 (721 mg, 68% yield) as a white solid. mp 155–156 °C; 1H NMR (400 MHz, CDCl3): δ 9.17 (s, 1H), 8.57 (s, 1H), 8.36 (br s, 1H), 8.22 (d, J = 8.2 Hz, 1H), 8.12 (s, 1H), 7.49–7.39 (m, 1H), 7.31–7.27 (m, 1H), 6.60 (d, J = 9.1 Hz, 1H), 3.59 (s, 4H), 3.25–3.15 (m, 2H), 2.74 (d, J = 5.0 Hz, 4H). Anal. (C16H18ClN5O·4HCl·H2O·1/2C3H8O) C, H, N.

2-(4-(Pyridin-2-yl)piperazin-1-yl)-N-(pyrimidin-5-yl)acetamide (27; CAB02-029HP)

Compound 27 was synthesized as described for 6 using K2CO3 (1.45 g, 10.5 mmol), 1-(pyridin-2-yl)piperazine (285 mg, 1.75 mmol), and compound 24b 2-chloro-N-(pyrimidin-5-yl)acetamide (300 mg, 1.75 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 27 (268 mg, 51% yield) as a white solid. mp 144–146 °C; 1H NMR (400 MHz, CDCl3): δ 9.17 (s, 1H), 9.02–8.81 (m, 3H), 8.13 (br s, 1H), 7.43 (br s, 1H), 6.60 (br s, 2H), 3.56 (br s, 4H), 3.18 (br s, 2H), 2.69 (br s, 4H). Anal. (C15H18N6O·3/2C2H2O4·3/4H2O) C, H, N.

2-(4-(5-Chloropyridin-2-yl)piperazin-1-yl)-N-(pyrimidin-5-yl)acetamide (28; CAB02-031HP)

Compound 28 was synthesized as described for 6 using K2CO3 (866 mg, 6.26 mmol), 1-(5-chloropyridin-2-yl)piperazine (207 mg, 1.04 mmol), and compound 24b 2-chloro-N-(pyrimidin-5-yl)acetamide (180 mg, 1.04 mmol) in an anhydrous acetonitrile (6 mL) solution. The crude product was purified by column chromatography (20–80% EtOAc/hexanes gradient) to give pure product 28 (157 mg, 45% yield) as a yellowish oil; 1H NMR (400 MHz, CDCl3): δ 9.21 (s, 1H), 9.08–8.98 (m, 3H), 8.15 (d, J = 3.7 Hz, 1H), 7.47 (d, J = 8.9 Hz, 1H), 6.63 (d, J = 9.2 Hz, 1H), 3.62 (br s, 4H), 3.30–3.23 (m, 2H), 2.77 (br s, 4H). Anal. (C15H17ClN6O·C2H2O4·1.5H2O) C, H, N.

Radioligand Binding Studies

HEK293 cells stably expressing human D2LR, D3R, or D4.4R were grown in a 50:50 mix of Ham’s F12 and Dulbecco’s modified Eagle’s medium culture media, supplemented with 2 mM l-glutamine, 20 mM N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid (HEPES), 0.1 mM non-essential amino acids, 1× antibiotic/antimycotic, 10% heat-inactivated fetal bovine serum, and 200 μg/mL hygromycin (Life Technologies, Grand Island, NY) and grown in an incubator at 37 °C and 5% CO2. Upon reaching 80–90% confluence, cells were harvested using pre-mixed Earle’s balanced salt solution (EBSS) 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 cell pellet was resuspended in 10 mL hypotonic lysis buffer (5 mM Tris, 5 mM MgCl2, pH 7.4 at 4 °C) and then centrifuged at 20000 rpm for 30 min at 4 °C. The membrane pellet was resuspended in fresh binding buffer for either [3H]N-methylspiperone (PerkinElmer, Waltham, MA) binding experiments [fresh EBSS buffer made from 8.7 g/L Earle’s balanced salts without phenol red (US Biological, Salem, MA), 2.2 g/L sodium bicarbonate, pH 7.4] or [3H]-(R)-(+)-7-OH-DPAT (ARC, Saint Louis, MO) binding experiments (50 mM Tris, 10 mM MgCl2, 1 mM EDTA, pH 7.4). A Bradford protein assay (Bio-Rad, Hercules, CA) was used to determine the protein concentration. Membranes were used fresh for [3H]-(R)-(+)-7-OH-DPAT binding experiments or diluted to 500 μg/mL and stored in a −80 °C freezer for later use in [3H]N-methylspiperone binding experiments.

All test compounds were freshly dissolved in 30% dimethyl sulfoxide (DMSO) and 70% H2O to a stock concentration of 100 μM. To assist the solubilization of free-base compounds, 10 μL of glacial acetic acid was added along with the DMSO. Each test compound was then diluted into half-log serial dilutions and tested in triplicate using the 30% DMSO vehicle. Competitive-inhibition experiments were conducted in 96-well plates containing 300 μL fresh binding buffer, 50 μL of diluted test compound, 100 μL of membrane suspension ([3H]N-methylspiperone: 20 μg/well for D2R and D3R, 30 μg/well for D4R; [3H]-(R)-(+)-7-OH-DPAT: 80 μg/well for D2R, 40 μg/well for D3R, 60 μg/well for D4R), and 50 μL of radioligand diluted in binding buffer ([3H]N-methylspiperone: 0.4 nM final concentration for all receptors; [3H]-(R)-(+)-7-OH-DPAT: 1.5 nM final concentration for D2R, 0.5 nM final concentration for D3R, 3 nM final concentration for D4R). Aliquots of [3H]N-methylspiperone and [3H]-(R)-(+)-7-OH-DPAT solution were also quantified accurately to determine how much radioactivity was added. Nonspecific binding was determined using 10 μM (+)-butaclamol (Sigma-Aldrich, St. Louis, MO), and total binding was determined with 30% DMSO vehicle. The reaction was incubated for 60 ([3H]N-methylspiperone) or 90 min ([3H]-(R)-(+)-7-OH-DPAT) at room temperature and terminated by filtration through PerkinElmer UniFilter-96 GF/B plates, presoaked in 0.5% polyethylenimine, using a Brandel 96-well plate harvester manifold (Brandel Instruments, Gaithersburg, MD). Filters were washed three times (∼1 mL/well) with ice cold binding buffer. After drying, 65 μL PerkinElmer MicroScint 20 scintillation cocktail was added to each well and filters were counted after at least 18 h of incubation using a PerkinElmer MicroBeta Microplate Counter. IC50 values for each compound were determined from dose–response curves and Ki values were calculated using the Cheng–Prusoff equation in GraphPad Prism 6 (GraphPad Software, San Diego, CA).[50]Kd values for [3H]N-methylspiperone and [3H]-(R)-(+)-7-OH-DPAT were determined via separate homologous competitive binding experiments at each receptor. Ki values for each compound/receptor/radioligand combination were calculated from at least three independent experiments and are reported as means ± SEM.

Functional Assays

β-Arrestin Recruitment Assay

Assays were conducted with minor modifications as previously published by our laboratory,[31,51−54] using the DiscoverX PathHunter technology (DiscoverX, Inc., Fremont, CA). Briefly, CHO-K1 cells stably expressing the human D2R long isoform, D3R, or D4R (DiscoverX, Inc.), were maintained in Ham’s F12 media supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, and 800 μg/mL G418 and 300 μg/mL hygromycin at 37 °C, 5% CO2, and 90% humidity. The cells were seeded in this media at a density of 2625 cells/well in 384-well black, clear-bottom plates. Compounds were diluted in phosphate-buffered saline in the presence of 0.2 μM sodium metabisulfite. Following overnight incubation, the cells were treated with multiple concentrations of compound and incubated at 37 °C for 90 min. DiscoverX reagent was then added to cells according to the manufacturer’s recommendations followed by 45–60 min incubation at room temperature. Luminescence was measured on a Hamamatsu FDSS μCell reader. Data were collected as RLUs and subsequently normalized to a percentage of the control luminescence seen with a maximum concentration of dopamine for agonist mode assays and the EC80 of dopamine for antagonist mode assays. The Hill coefficients of the concentration–response curves did not significantly differ from unity.

cAMP Inhibition Assay

D4R- and D2R-mediated inhibition of forskolin-stimulated cAMP production was assayed using the PerkinElmer LANCE Ultra cAMP assay kit (PerkinElmer, Inc., Waltham, MA). CHO-K1 cells stably expressing the human D2R long isoform or D4R were maintained in Ham’s F12 supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, and 800 μg/mL G418 and 300 μg/mL hygromycin at 37 °C, 5% CO2, and 90% humidity. Cells were seeded in Hank’s balanced salt solution (with CaCl and MgCl2) with 5 mM HEPES buffer and 0.2 μM sodium metabisulfite at a density of 5000 cells/well in 384-well white plates. Compounds and forskolin were made in the same buffer. Immediately after plating, cells were treated with 2.5 μL of compound (at various concentrations) and 2.5 μL of forskolin and incubated at room temperature for 30 min. The final concentration of forskolin was 10 μM. When running assay in antagonist mode, the EC80 of dopamine (10 nM) was added with the forskolin solution. Eu-cAMP tracer and ULight-anti-cAMP solutions were added as directed by the manufacturer and cells were incubated for 2 h in the dark at room temperature, after which a time-resolved fluorescence resonance energy transfer (TR-FRET) signal was measured using a BMG Labtech PHERAstar Fs (BMG Labtech USA, Cary, NC). Values were normalized to a percentage of the control TR-FRET signal seen with a maximum concentration of dopamine for agonist mode assays and the EC80 of dopamine for antagonist mode assays. The Hill coefficients of the concentration–response curves did not significantly differ from unity with the data fitting to a single site model.

Molecular Docking Studies

Crystal Structures of D2R, D3R, and D4R

In this study, we used the crystal structure of the human dopamine D2, D3, and D4 receptor in complex with antagonists risperidone (PDB: 6CM4(26)), eticlopride (PDB: 3PBL(25)), and nemonapride (PDB: 5WIU(24)), respectively. Each of the three crystal structures was prealigned in membrane using the OPM web server.[55]

Protein Structure Preparation

The structures of D2R, D3R, and D4R were further prepared using Maestro Protein Preparation Wizard.[56] First, the hydrogens and missing side chains were added. Second, the protonation state of the receptor was optimized at pH = 7. Third, a restrained minimization was performed to relax the receptor structure using OPLS3 force field.[57]

Ligand Preparation

The 2D structures of 1, 6, 9, 10, 12, 13, and 21 were first constructed in ChemDraw and then converted into a 3D structure using Maestro Elements. Next, the protonation state was generated at pH = 7 using the pKa prediction program Epik that is based on the Hammett and Taft methodologies.[56] Lastly, the geometry of each ligand was optimized using an energy minimization.

Ligand Docking

The orthosteric ligand pockets of D2R, D3R, and D4R were specified by the crystal ligands risperidone, eticlopride, and nemonapride, respectively, and a 3D box was formed around each crystal ligand to enclose the orthosteric ligand binding pocket. Each ligand was first docked using the Glide XP scoring function with default procedures and parameters.[58,59] Reproductions of the crystal binding poses of risperidone, eticlopride, and nemonapride in D2R, D3R, and D4R, respectively, provide a solid validation for our XP docking protocol (Figures S2–S4). To refine the docking poses of noncrystal ligands, induced fit docking (IFD) was conducted on the complex from the Glide XP docking.

MD Simulation System Setup

Seven MD simulation systems were built using the complexes from the IFD. Each pre-aligned complex was placed in a double lipid membrane formed by 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine lipids[60] and then solvated in an orthorhombic water box with a buffer distance of 10 Å using the SPC water model.[61] Each system was neutralized using Na+ ions, added with a salt concentration of 0.15 M NaCl. The OPLS3 force field[57] was used to represent the receptor–ligand–lipid system.

Relaxation and Production Runs

Using Desmond, each system was first relaxed using the default relaxation protocol for membrane proteins.[62] After the relaxation, a 100.0 ns production run was conducted under the NPT ensemble for each system using the default protocol. A temperature of 300 K was controlled using the Nosé–Hoover chain coupling scheme[63] with a coupling constant of 1.0 ps. A pressure of 1 atm was controlled using the Martyna–Tuckerman–Klein chain coupling scheme[63] with a coupling constant of 2.0 ps. M-SHAKE[64] was applied to constrain all bonds connecting hydrogen atoms, enabling a 2.0 fs time step in the simulations. The k-space Gaussian split Ewald method[65] was used to treat long-range electrostatic interactions under periodic boundary conditions (charge grid spacing of ∼1.0 Å, and direct sum tolerance of 10–9). The cutoff distance for short-range nonbonded interactions was 9 Å, with the long-range van der Waals interactions based on a uniform density approximation. To reduce the computation, nonbonded forces were calculated using an r-RESPA integrator[66] where the short-range forces were updated every step and the long-range forces were updated every three steps. The trajectories were saved at 100.0 ps intervals for analysis.

SID Analyses

The SID tool was used to generate graphical information about the behavior and interaction of the protein and ligand during simulation. The analysis gives us graphical representation of root mean square deviation (RMSD), root mean square fluctuation, secondary structures changes, protein–ligand contacts, and ligand torsion profiles of rotatable bonds.

Convergence of Simulations

To check the convergence of the simulations, we investigated the protein Cα and ligand RMSD plots for each system (Figures S34–S36). The relatively flat plots within last 20 ns indicate that the complex systems have reached a steady state.

Trajectory Clustering Analyses

The Desmond trajectory clustering tool[67] was used to group complex structures for each system. The backbone RMSD matrix was used as structural similarity metric and hierarchical clustering with average linkage[67] was selected as the clustering method. The merging distance cutoff was set to be 2.5 Å. For all systems, a dominant cluster with was identified to have more than 80% of the trajectory population.