The Molecular Basis of Species-Specific Ligand Activation of Trace Amine-Associated Receptor 1 (TAAR₁)

Abstract

The trace amine-associated receptor 1 (TAAR₁) is an aminergic G protein-coupled receptor (GPCR) potently activated by 3-iodothyronamine (1), an endogenous derivative of thyroid hormone. Structure activity relationship studies on 1 and related agonists showed that the rat and mouse species of TAAR₁ accommodated structural modifications and functional groups on the ethylamine portion and the biaryl ether moiety of the molecule. However, the two receptors clearly exhibited distinct, species-specific ligand preferences despite being remarkably similar with 93% sequence similarity. In this study, we generated single and double mutants of rat and mouse TAAR₁ to probe the molecular recognition of agonists and the underlying basis for the ligand selectivity of rat and mouse TAAR₁. Key, non-conserved specificity determinant residues in transmembranes helices 4 and 7 within the ligand binding site appear to be the primary source of a number of the observed ligand preferences. Residue 7.39 in transmembrane 7 dictated the preference for a β-phenyl ring while residue 4.56 in transmembrane 4 was partially responsible for the lower potency of 1 and tyramine for the mouse receptor. Additionally, 1 and tyramine were found to have the same binding mode in rat TAAR₁ despite structure activity relationship data suggesting the possibility of each molecule having different binding orientations. These findings provide valuable insights into the critical binding site residues involved in the ligand-receptor interaction that can influence compound selectivity and functional activity of aminergic GPCRs.

INTRODUCTION

3-Iodothyronamine (1, T₁AM; Fig. 1) is an endogenous derivative of the thyroid hormone thyroxine (T₄; Fig. 1) detected in the brain, heart, liver, and blood (1, 2). It has profound pharmacological effects in vivo that are mostly opposite to those of T₄ (3). Compound 1 induces anergia, hypothermia, bradycardia, hyperglycemia, and hyperinsulinemia when administered to mice and rapidly reduces cardiac output in an ex vivo working rat heart preparation (2, 4–8). Additionally, 1 can increase food intake and influence energy metabolism by favoring lipid utilization over carbohydrate consumption (9, 10).

Figure 1.

Thyroid hormone, 3-iodothyronamine, and related analogs. Structures of thyroxine, 3-iodothyronamine (1), β-phenylphenoxyphenethylamines (2 and 3), and phenoxynaphethylamine (4). The A, B, and C rings of 2 and 3 correspond to the outer, inner, and β-phenyl rings, respectively.

These physiological responses to 1 may be mediated by more than one molecular target. In vitro, 1 can activate aminergic G protein-protein coupled receptors (GPCRs) in the biogenic amine subfamily; stimulating the production of cAMP (adenosine 3′,5′-cyclic monophosphate) via the trace amine-associated receptor 1 (TAAR₁) and the degradation of cAMP via the α _2A adrenergic receptor (2, 7, 11, 12). Additionally, 1 has a neuromodulatory role, inhibiting neurotransmitter reuptake by the dopamine (DAT) and norepinephrine transporter (NET), and inhibiting vesicular packaging by the vesicular monoamine transporter 2 (VMAT2) (13).

To date, 1 is the most potent endogenous molecule that can activate rat and mouse TAAR₁. TAAR₁ is a member of the trace amine-associated receptor family of orphan GPCRs, which consists of 19 rat, 16 mouse, and 9 human subtypes (14–18). It is homologous to the dopamine, adrenergic, and serotonin receptors and expressed in multiple tissues including the heart, kidney, liver, spleen, and pancreas. Although a biological function for any TAAR subtype has yet to be defined, rodent and human species of TAAR₁ are potently activated by endogenous biogenic amines such as β-phenethylamine as well as the psychostimulants amphetamine and methamphetamine in a species-dependent manner (12, 14, 15, 19–21). In addition, other members of the rodent TAAR family were recently implicated as a secondary class of chemosensory receptors expressed in the olfactory epithelium (22).

In an effort to determine the role of TAAR₁ in mediating the effects of 1, we have previously explored the structure activity relationship (SAR) of 1 and developed a number of analogs with improved agonist activity (11, 23). These SAR studies showed rat and mouse TAAR₁ (rTAAR₁ and mTAAR₁, respectively) could tolerate structural modifications and substituents on the ethylamine portion and biaryl ether moiety of 1. However the two receptors clearly have distinct structural preferences. In the ethylamine chain, for example, rTAAR₁ favored unsaturated hydrocarbon substituents while mTAAR₁ preferred polar groups and hydrogen bond acceptors (23). Compound 1 and tyramine (Table 1), an endogenous phenethylamine metabolite, also exhibited some degree of species variability, being ~7 to 10-fold less potent for mTAAR₁ compared to rTAAR₁ (11). These distinct ligand selectivities of rat and mouse TAAR₁ are unusual given that the two receptors are 93% similar in primary sequence. Since the rodent receptors are only 83%–85% similar to human TAAR₁, understanding the molecular basis of species variability will undoubtedly have important implications in the development of activators and inhibitors for human TAAR₁. In this study, we explored the molecular recognition of 1 and related agonists by TAAR₁ and identified specificity determining residues that give rise to the disparate ligand preferences between rat and mouse TAAR₁.

Table 1.

Agonist activity of 1–12 and tyramine on wild type rat and mouse TAAR₁

						rTAAR1			mTAAR1
						EC50a	Emaxb	Nc	EC50a	Emaxb	Nc
Compd	R1	R2	R3	R4	R5	± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)
1 (T1AM)	p- OH - Ph	I	H	H	-	33± 3	100± 0	5	314 ± 43	100 ± 0	5
2 (ET-13)	OPh	H	Ph	H	-	28 ± 2	103 ± 4	3	>10,000	35 ± 8	3
3 (ET-14)	H	OPh	Ph	H	-	19 ± 2	131 ± 7	3	>10,000	15 ± 4	3
5 (ET-36)	OPh	H	p- OH- Ph	H	-	6 ± 1	114 ± 9	4	>10,000	62 ± 6	3
6 (ET-64)	OPh	H	p- OH- Ph	CH3	-	5 ± 1	127 ± 2	4	>10,000	42 ± 1	2
7 (ET-69)	p- OH - Ph	I	p- OH- Ph	H	-	4 ± 1	115 ± 2	6	>10,000	34 ± 5	3
8 (ET-71)	OH	H	Ph	H	-	78 ± 9	122 ± 16	3	>10,000	49	1
9 (ET-50)	OH	H	p- OH- Ph	H	-	115 ± 12	105 ± 5	3	>10,000	72	1
10 (PTA)	OPh	H	H	H	-	63 ± 7	93 ± 4	3	420 ± 66	85 ± 4	3
Tyramine	OH	H	H	H	-	65 ± 1	119 ± 7	3	271 ± 52	110 ± 2	3
4 (ET-21)	-	-	-	-	Ph	26 ± 1	113 ± 5	3	100 ± 22	104 ± 3	3
11 (ET-102)	-	-	-	-	p- OH-Ph	19 ± 3	96 ± 2	3	171 ± 13	98 ± 1	2
12 (1-NEA)	-	-	-	-	H	65 ± 6	115 ± 2	3	82 ± 17	112 ± 3	2

^aEC₅₀ is the half-maximal effective concentration of a compound.

^bE_max is the maximum stimulation achieved at a concentration of 10 mM and was calculated by use of Prism software. EC₅₀ and E_max values represent the average of N independent experiments in triplicate and were calculated by use of Prism software as escribed in the Methods section. E_max = 100 % is defined as the activity of 1 at 10 mM.

^cN is the number of independent experiments in triplicate that were performed and used to calculate the EC₅₀ and E_max values.

RESULTS AND DISCUSSION

Our initial structure activity relationship study identified the β-phenylphenoxy-phenethylamine (2 & 3) and phenoxynaphethylamine (4) as promising scaffolds for the development of small molecule regulators of TAAR₁ (Fig. 1) (23). Compounds 2 and 3 were both rTAAR₁ selective, being 357- to 526-fold more potent for rTAAR₁ compared to mTAAR₁ (Table 1). Compound 4, on the other hand, was less discriminating between rat and mouse TAAR₁ with a disparity in potency of 5-fold between the two receptors. Using the toggle switch model of aminergic GPCR activation as a guideline, 2 was further developed to give superagonists (agonists that are more potent and/or efficacious than 1) 5, 6, and 7 (Table 1) (24). These superagonists remained rTAAR₁ selective exhibiting potencies that were ≥ 1600- to 2500-fold better at rTAAR₁ versus mTAAR₁.

The poor agonist activity of the β-phenylphenoxyphenethylamines at mTAAR₁ can be attributed to the β-phenyl group of the molecule and not its outer ring moiety (Fig. 1). Removing the outer rings of 2 and 5 (8 and 9, respectively) did not increase potency at mTAAR₁ (Table 1). Both 8 and 9 still activated mTAAR₁ poorly with a potency ≥10μM. In contrast, eliminating the β-phenyl ring of 2, 5, and 7 to give 1 or 10 improved the potency at mTAAR₁ ~24- to 32-fold and decreased the rTAAR₁ selectivity from ≥ 350- to 2500-fold down to ~7–10-fold.

Aminergic GPCRs are heptahelical transmembrane proteins with an extracellular amino terminus and an intracellular carboxy terminus (Fig. 2a). The ligand binding site of aminergic GPCRs is located within the transmembrane (TM) region of the receptor and is predominantly composed of residues from TM 3, 5, 6 and 7 (25–28). Based on pharmacological and mutagenesis studies, epinephrine is proposed to bind to the β₂-adrenergic receptor (β₂AR) with aspartic acid 3.32 (D3.32) acting as the counterion for the charged amine, serine residues 5.42, 5.43, and 5.46 (S5.42, S5.43, and S5.46, respectively) interacting with the catechol hydroxyls, phenylalanines 6.51 and 6.52 (F6.51 and F6.52) interacting with the catechol ring, and asparagines 6.55 (N6.55) as the partner for the β-hydroxy group (Fig. 2b) (see experimental procedures for a description of the residue indexing system) (29–36). In the recently solved crystal structure of an engineered human β₂AR, all of these residues were found to line the ligand binding site and interact with the inverse agonist that co-crystallized with the receptor (26, 27). By analogy to the catecholamine receptors (dopamine, epinephrine, and norepinephrine) we have previously deduced 2 to bind to rTAAR₁ with the charged amine forming a salt bridge interaction with D3.32 and the biaryl ether oxygen hydrogen bonding to S5.46 (Fig. 2c) (24). In this binding orientation, our homology model of rTAAR₁ showed the β-phenyl ring to be positioned near the interface between TM 6 and 7 and surrounded by cysteine 6.54, methionine 6.55, and asparagines 7.35 and 7.39 (C6.54, M6.55, N7.35, and N7.39, respectively).

Figure 2.

Biogenic amine GPCRs. (a) Schematic representation of the helical arrangement of GPCRs viewed from the cell membrane. (b) Binding orientation of (R)-epinephrine in the binding site of the β₂AR. (c) Proposed binding orientation of 2 in the binding site of rTAAR₁. (d) Binding site of mTAAR₁. (e) Binding site of hTAAR₁. The binding sites of β₂AR, rTAAR₁, and mTAAR₁ are viewed from the perspective TM4. The rotamer switch residues (white letters), proposed binding and specificity determinant residues are labeled. Non-conserved residues that were mutated are shown in red.

In mTAAR₁, C6.54 and N7.35 are conserved but not M6.55 and N7.39. Instead mTAAR₁ has a threonine and tyrosine at 6.55 and 7.39, respectively (T6.55 and Y7.39) (Fig. 2d). In the β₂AR, the residues at 6.55 and 7.39 have been previously shown to interact with the ligand and alter the receptor’s ligand specificity (35, 37). Since the β-phenyl ring is proposed to be in the vicinity of 6.55 and 7.39, and both residues are not conserved between rat and mouse TAAR₁, we hypothesized that one or both of these residues are specificity elements that influence compatibility with the β-phenyl ring of the β-phenylphenoxyphenethylamine scaffold. We tested this hypothesis by measuring the activity of 3 and 5, representative β-phenylphenoxyphenethylamines, against rat and mouse TAAR₁ in HEK293 (human embryonic kidney 293) cells stably and heterologously expressing single or double swap mutants at 6.55 and/or 7.39.

Residue 7.39 controls specificity for the β-phenyl ring of TAAR₁ ligands

Swapping residue 6.55 of rat and mouse TAAR₁ had minor effects on the activity and selectivity of 3 and 5. In the rTAAR₁ 6.55 single mutant [rTAAR₁(M6.55T)] the potency and efficacy of 3 (EC₅₀ = 55 ± 20 nM, E_max = 119 ± 6%) and 5 (EC₅₀ = 11 ± 1 nM, E_max = 117 ± 20%) decreased 2- to 3-fold and ≤ 12%, respectively (Table 2). When T6.55 was mutated to a methionine in mTAAR₁ [mTAAR₁(T6.55M)], both compounds were still poor agonists activating the receptor with potencies >10μM and efficacies ranging from 20% to 45%. Despite the mutation at residue 6.55, both 3 and 5 remained considerably rTAAR₁ selective (182- to 909-fold) (Fig. 3).

Table 2.

Agonist activity of 1, 3–5, and 11 on rTAAR₁ TM 6 and/or 7 mutants

	rTAAR1
	M6.55T			N7.39Y			M6.55T/N7.39Y
	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc
Compd	± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)
1	129 ± 43	100 ±0	2	135	100	1	90	100	1
3	55 ± 20	119 ± 6	2	~10,000	85	1	>10,000	46	1
4	59 ± 0	117 ± 0	2	146	129	1	75	111	1
5	11 ± 1	117 ± 20	2	~1,000	126	1	~2,000	70	1
	mTAAR1
	T6.55M			Y7.39N			T6.55M/Y7.39N
	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc
Compd	± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)
1	1418 ± 592	100 ± 0	2	251 ± 43	100 ± 0	5	208 ± 60	100 ± 0	3
3	>10,000	20 ± 4	2	102 ± 21	95 ± 6	5	43 ± 5	110 ± 8	3
4	350 ± 101	96 ± 13	2	173 ± 41	93 ± 15	2	-	-	-
5	>10,000	45 ± 5	2	176 ± 32	91 ± 11	5	138 ± 24	101 ± 9	3
11	-	-	-	179 ± 21	92 ± 6	3	134 ± 43	101 ± 11	3

^a–cSee footnotes for Table 1. Compound structures are shown in Table 1.

Figure 3.

Selectivity profiles of 3 and 5 on rat and mouse TAAR₁ wild type and mutant receptors. The top and bottom panels show the potency and efficacy, respectively, of 3 and 5. The EC₅₀ and E_max for rTAAR₁ and mTAAR₁ receptors are depicted in solid and hollow symbols, respectively. WT rTAAR₁ [■], rTAAR₁(M6.55T) [], rTAAR₁(N7.39Y) [], rTAAR₁(M6.55T/N7.39Y) [], WT mTAAR₁ [□], mTAAR₁(T6.55M) [], mTAAR₁(Y7.39N) [], and mTAAR₁(T6.55M/Y7.39N) []. The lines connecting the solid and hollow symbols represent the difference in potency (top panels) and efficacy (bottom panels) between rat and mouse TAAR1 receptors. The fold difference in potency and percent difference in efficacy are listed above (top panels) and to the left (bottom panels) of the connecting lines, respectively.

The single swap mutants at residue 7.39 had opposing effects on the activity of 3 and 5. The potency of both compounds decreased 167- to 526-fold in the rTAAR₁ 7.39 mutant [rTAAR₁(N7.39Y)] (EC₅₀ = ~10 μM and ~1 μM for 3 and 5, respectively) but increased 60- to 100-fold in its mTAAR₁ 7.39 mutant counterpart [mTAAR₁(Y7.39N)] (EC₅₀ = 102 ± 21 nM and 176 ± 32 nM for 3 and 5, respectively) (Table 2). The efficacy of 3 (E_max = 95 ± 6%) and 5 (E_max = 91 ± 11%) both increased 29% to 89% in mTAAR₁(Y7.39N). For rTAAR₁(N7.39Y) the efficacy of 5 (E_max = 126%) increased 12% while that of 3 (E_max = 85%) decreased 46% compared to wild type. Interestingly, swapping the residues at 7.39 converted both compounds to good mTAAR₁ agonists that were now 6- to 98-fold selective for mTAAR₁ over rTAAR₁ (Fig. 3).

The activity profile of 3 and 5 for the double swap mutants was similar to that of the 7.39 single mutants but showed some enhancements in both potency and efficacy. The potency of 3 and 5 decreased 333- to 526-fold for the rTAAR₁ 6.55 and 7.39 double mutant [rTAAR₁(M6.55T/N7.39Y)] but increased 70- to 320-fold for the mTAAR₁ 6.55 and 7.39 double mutant equivalent [mTAAR₁(T6.55M/Y7.39N)] (Table 2). The same trend was also observed with regards to efficacy; decreasing 44% to 85% for rTAAR₁(M6.55T/N7.39Y) and increasing 39% to 95% for mTAAR₁(T6.55M/Y7.39N). Compared to the 7.39 single mutant, the mTAAR₁ selectivity of 3 and 5 in the double mutant was more pronounced (Fig. 3).

The decrease in activity of 3 and 5 for the 6.55 and/or 7.39 single and double mutants in rTAAR₁ cannot be attributed to the introduced mutations compromising the functional competency of the receptors because the activity of the positive controls (1 and 4) for the same mutants only changed ≤ 4-fold and ≤ 16% in terms of potency and efficacy, respectively (Table 2). Likewise, the enhanced activity of 3 and 5 for the mTAAR₁ single and double swap mutants is not a consequence of the mutations rendering the receptors constitutively active and more responsive to agonists because the potency of the positive controls (1, 4, and/or 11) only changed ≤ 2-fold and the efficacy were essentially identical compared to the wild type receptor (Table 1 and 2). Compound 11 is a novel agonist for rat and mouse TAAR₁ that can be considered a halogen free analog of 1.

Swapping residues at 7.39 was sufficient to convert 3 and 5 from a rTAAR₁ into a mTAAR₁ selective agonist. The TAAR₁ binding site appears to be able to accommodate a phenyl ring in the interface between TM 6 and 7 within the binding site near residue 7.39. Compounds 3 and 5 are poor agonists for wild type mTAAR₁ because this phenyl pocket near 7.39 is occupied by the phenol group of Y7.39 from the receptor and unavailable to the β-phenyl rings of 3 and 5. In contrast, 3 and 5 are excellent agonists for wild type rTAAR₁ because the smaller asparagine residue at 7.39 is less sterically encumbering and does not compete with the β-phenyl rings of 3 and 5 for the phenyl pocket near 7.39 of the receptor.

It should be noted that the substitution pattern of the outer ring does not significantly affect the specificity of TAAR₁ for β-phenyl ring bearing compounds. Regardless of whether the outer ring is at the meta- (3) or para- (5) position relative to the ethylamine chain, both 3 and 5 were affected similarly by mutations at 6.55 and/or 7.39 in rat and mouse TAAR₁, indicating that the β-phenyl ring of both molecules occupy the same binding pocket within the binding site. This result supports the assumptions we had proposed regarding the location of the antagonistic groups of the lead rTAAR₁ antagonists previously developed with 3 as the core scaffold (24).

Based on its primary sequence, we predict human TAAR₁ (hTAAR₁) to be able to accommodate a phenyl ring at the β carbon of phenethylamine based ligands because the residue at 7.39 in the human receptor is an isoleucine (Fig. 2e). Since isoleucine is smaller than tyrosine and approximately the same size as asparagine, the phenyl pocket near 7.39 should also be present in hTAAR₁.

Compound 1 and Tyramine have similar binding modes

In addition to being a superagonist for rTAAR₁, 7 (EC₅₀ = 4 ± 1 nM, E_max = 115 ± 2%) is also an interesting molecule because it contains the structures of both tyramine (EC₅₀ = 65 ± 1 nM, E_max = 119 ± 7%) and 1 (EC₅₀ = 33 ± 3 nM, E_max = 100 ± 0%) in the same compound (Table 1 and Fig. 4). This hybrid compound potentially explains how two molecules with very different molecular volumes can elicit similar responses. If the β-phenyl group of 7 represents the aromatic ring of tyramine, then tyramine would occupy the phenyl pocket near 7.39 and thus have a binding mode different from 1 (Fig. 4). On the other hand, if the inner ring of 7 represents the tyramine aromatic ring then tyramine and 1 would have similar binding modes. The rTAAR₁ agonist activity of 9 (EC₅₀ = 115 ± 12 nM, E_max = 105 ± 5%) supports the feasibility of tyramine potentially having two alternate binding modes in the rat receptor (Table 1). Since the tyrosine residue at 7.39 in mTAAR₁ abolished the phenyl pocket near 7.39, tyramine must share the same binding mode as 1 in mTAAR₁.

Figure 4.

Proposed binding modes of 7, 1, and tyramine in rTAAR₁. The binding site of rTAAR₁ is viewed from the perspective of TM4 (see Fig. 2a). The rotamer switch residues (white letters), proposed binding and specificity determinant residues are labeled. Residue A5.42 is shown in red. The similar and distinct binding mode models of 1 and tyramine are outlined in green and blue dashed lines, respectively.

To determine if tyramine and 1 have similar or distinct binding modes, we mutated alanine 5.42 in rTAAR₁ to a threonine, leucine, or isoleucine [rTAAR₁(A5.42T), rTAAR₁(A5.42L), and rTAAR₁(A5.42I), respectively] and examined the effects of the mutations on the potency of tyramine and 1. The idea behind these 5.42 single mutants is to perturb the pocket occupied by the outer ring of 1, 7 and other biaryl ether containing molecules by introducing larger residues than alanine. Threonine was chosen because SAR studies on human TAAR₁, which has a threonine at residue 5.42 (Fig. 2e), showed 1 to be a less potent agonist against hTAAR₁ than rTAAR₁ (12).

In the rTAAR₁(A5.42L) mutant, the potency of 1 (EC₅₀ = 58 ± 16 nM) and tyramine (EC₅₀ = ~2 μM) decreased 2- and 31- fold, respectively (Table 3a). When A5.42 was mutated to isoleucine, there was a 3- and 154-fold decrease in the potency of 1 (EC₅₀ = 108 ± 14 nM) and tyramine (EC₅₀ = >10 μM), respectively. The efficacy of tyramine in the leucine mutant decreased 53% to 70% compared to wild type rTAAR₁. A similar trend was observed for the rTAAR₁(A5.42T) mutant where 1 (EC₅₀ = 88 ± 13 nM) and tyramine (EC₅₀ = >10 μM) were 3- and 154-fold less potent. Additionally, tyramine was 50% less efficacious for this receptor than wild type. Since the activity of the positive control (11) for all 5.42 mutants were essentially unaffected, the reduced activities of 1 and tyramine for these mutants cannot be due to a compromised activation capacity of these receptors.

Table 3.

Agonist activity of 1, 4, 11, and tyramine on TAAR₁ TM 4 or 5 mutants

a. TAAR₁ TM 5 mutants

	rTAAR1
	A5.42L			A5.42I			A5.42T
	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc
Compd	± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)
1	58 ± 16	100 ± 0	3	108 ± 14	100 ± 0	3	88 ± 13	100 ± 0	4
Tyramine	~2,000	102 ± 18	3	>10,000	49 ± 7	3	>10,000	69 ± 5	4
11	32 ± 12	101 ± 8	3	29 ± 4	94 ± 2	3	25 ± 3	117 ± 13	4

b. TAAR1 TM 4 mutants
	rTAAR1			rTAAR1
	A5.42L			A5.42I
	EC50a	Emaxb	Nc	EC50a	Emaxb	Nc
Compd	± SEM (nM)	± SEM (%)		± SEM (nM)	± SEM (%)
1	38 ± 11	100 ± 0	4	67 ± 17	100 ± 0	6
Tyramine	54 ± 11	108 ± 6	4	40 ± 13	118 ± 8	6
4	22 ± 5	101 ± 14	4	100 ± 16	123 ± 7	6
11	27 ± 15	101 ± 6	4	123 ± 41	105 ± 5	6

^a–cSee footnotes for Table 1. Compound structures are shown in Table 1.

The observed effects on the activity of 1 and tyramine in the rTAAR₁ 5.42 mutants are consistent with both compounds having similar binding modes. If tyramine preferentially occupied the phenyl pocket near residue 7.39, its potency would not have been affected by changes to residue 5.42. Since its activity decreased 31- to 154-fold when residue 5.42 was mutated, tyramine is probably binding to rTAAR₁ with its hydroxyl group engaged in hydrogen bond interactions with S5.46, the residue one turn below 5.42 in TM5 (Fig. 2d). This binding mode corresponds to the same binding orientation of 1 for rat and mouse TAAR₁.

Although the leucine and isoleucine mutations increased the steric bulk around residue 5.42, it is interesting that the agonist activity of the smaller tyramine was more severely affected than the larger 1 or 11; especially when these mutations were intended to block the binding pocket for the outer ring. These results suggest that despite being bigger than alanine, leucine and isoleucine are not large enough to completely abolish the outer ring binding pocket or are sufficiently flexible to accommodate the outer phenyl ring. The small effects of the mutations on the agonist activity of 1 and 11 can potentially be attributed to their larger number of interactions with the receptor compared to tyramine. Like tyramine, 1 and 11 should both be anchored in the binding site of rTAAR₁ by a salt bridge interaction between the charged amine and D3.32, and a hydrogen bond interaction with S5.46. However, the extra functional groups present in 1 and 11 (i.e. outer ring, iodine, and naphthyl ring) (Table 1) can make additional interactions with TM 5 and 6 that are not available to tyramine. With more contacts to the receptor, 1 and 11 would be less sensitive to structural changes at residue 5.42 than tyramine.

Residue 4.56 is partially responsible for the lower potency of 1 for mTAAR₁

Within the thyronamine series, rat and mouse TAAR₁ had the same rank order potency but the potency values of individual compounds for the two receptors were not identical. In general, thyronamines are ~10-fold less potent for mTAAR₁ compared to rTAAR₁ (2, 11). Assuming all ligands target the same binding pocket and lead to the same active receptor conformation, a possible explanation for this potency disparity can be attributed to a lower G-protein coupling efficiency for mTAAR₁ versus rTAAR₁. If this were the case, then it would be impossible to have an equipotent agonist for both receptors because mTAAR₁ would be inherently less sensitive to ligand activation than rTAAR₁. Since 12 was found to be an equipotent agonist for rTAAR₁ (EC₅₀ = 65 ± 6 nM, E_max = 115 ± 2%) and mTAAR₁ (EC₅₀ = 82 ± 17 nM, E_max = 112 ± 3%), the G-protein coupling efficiency of mTAAR₁ is comparable to that of rTAAR₁ (Table 1).

We hypothesized that the potency disparity of thyronamines was brought about by nonconserved amino acid(s) at key specificity determinant residues within the binding site. In particular, we speculated that tyrosine 4.56 (Y4.56) was primarily responsible for the ~10-fold lower potency of 1 for mTAAR₁. This residue was identified through a process of elimination based on the following five points: (1) since the binding sites for most aminergic GPCRs are located within the transmembrane regions of the receptor, all intracellular and extracellular loops as well as the amino- and carboxy-terminus were eliminated (28); (2) amino acid differences in TM 1 and 2 were eliminated because the binding site is primarily composed to TM 3, 4, 5, 6, and 7 (30); (3) since the ethylamine chain of 1 and 12 are exactly the same, non-conserved residues in TM 3, 6 and 7 cannot be responsible; (4) TM5 was eliminated because it is absolutely conserved between the two species; (5) the intracellular half of TM 4 was eliminated because the binding site of GPCRs is located in the extracellular half of the transmembrane region (Fig. 5). The only non-conserved residue remaining was Y4.56. In our homology models of rat and mouse TAAR₁, residue 4.56 was found to be in the vicinity of the purported outer ring binding pocket and could conceivably make contacts with the bound ligand. To test the importance of residue 4.56, we generated rat and mouse TAAR₁ single swap mutants at this location.

Figure 5.

Sequence comparison of rat and mouse TAAR₁. Dots represent conserved residues. Amino and carboxy termini (N-terminus and C-terminus, respectively), intracellular loops (IL), extracellular loops (EL), and transmembrane regions (TM) are labeled. The most conserved residue in each TM region is labeled X.50 and residue 4.56 is highlighted in green.

When Y4.56 of mTAAR₁ was converted to phenylalanine [mTAAR₁(Y4.56F)], the potency of 1 increased 5-fold from 314 ± 43 nM to 67 ± 17 nM (Table 1 and 3b). Interestingly, the potency of 1 (EC₅₀ = 38 ± 11 nM) for the tyrosine mutant of rTAAR₁ [rTAAR₁(F4.56Y)] was comparable to wild type rTAAR₁ (33 ± 3 nM). The same trend was also observed for tyramine; where its potency increased 7-fold in mTAAR₁(Y4.56F) (EC₅₀ = 40 ± 13 nM) but remained unaffected in rTAAR₁(F4.56Y) (EC₅₀ = 54 ± 11 nM). Since the agonist activity of the positive controls (4 and 11) for the mutants was comparable to that of the wild type receptors for both species, the mutations did not compromise the functional capacity of either receptors.

Residue 4.56 appears to play an important role in the lower potency of 1 for mTAAR₁. Swapping this residue with the corresponding residue found in rTAAR₁ increased the potency of 1 in mTAAR₁ to within two-fold of the potency value for wild type rTAAR₁. Interestingly, the potency of tyramine for mTAAR₁(Y4.56F) was equivalent to that of wild type rTAAR₁. Although mutating residue 4.56 improved the potency in the mTAAR₁ mutants, the reciprocal effect of 1 and tyramine becoming less potent for rTAAR₁(F4.56Y) was not observed. This indicates that 4.56 is only partially responsible and not the sole basis for the observed potency disparity of 1 and tyramine between the rat and mouse TAAR₁. Overall, these results implicate residue 4.56 in TM4 to affect ligand interaction and receptor activation.

CONCLUSION AND SIGNIFICANCE

The disparate ligand structural preferences exhibited by rat and mouse TAAR₁ can be attributed to key, non-conserved specificity determining residues within the binding site. Residue 7.39 appears to dictate the specificity for a β-phenyl ring; the results suggest that the bulky tyrosine residue at 7.39 in mTAAR₁ sterically clashed with the β-phenyl ring whereas the smaller asparagine at the same location in rTAAR₁ was more compatible and able to accommodate a β-phenyl moiety. The lower potency of 1 in mTAAR₁ was partly caused by the presence of a tyrosine at residue 4.56 rather than a phenylalanine. Although compound 7 implied the possibility of 1 and tyramine having different binding modes in the binding site of rTAAR₁, 1 and tyramine appear to have the same binding orientation.

With the recent developments and accomplishments in structure determination of the β₂ adrenergic receptor, the practicality of a structure-based drug design approach towards developing activators and inhibitors for aminergic GPCRs has never been so promising. A critical aspect to the success of this strategy will depend on having insights into the molecular basis of ligand recognition, the mechanism of GPCR activation, and the relationship of how these ligand-receptor interactions are relayed and translated into receptor activation or inhibition. The information presented herein should prove beneficial towards this cause as it provides valuable information regarding the binding site residues involved in ligand-receptor interactions that can influence compound specificity and functional activity of an aminergic GPCR.

METHODS

Residue Indexing Scheme

Residues are labeled relative to the most conserved amino acid in the transmembrane segment in which it is located (38). Asparagine 7.39, for example, is located in transmembrane 7 and precedes the most conserved residue by 11 positions. Proline 6.50 is the most conserved residue in TM6. This system simplifies the identification of corresponding residues in different GPCRs.

TAAR₁ Site-Directed Mutagenesis

TAAR₁ mutants were generated by using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Primers were designed that coded for the desired mutation flanked with 10–15 base pairs of sequence. Complementary oligonucleotides were then used in PCR using an expression plasmid containing the desired receptor as template. The PCR product was digested with Dpn I and transformed into XL1 Blue competent cells. Colonies were picked and the DNA isolated was sequenced to confirm the mutation.

The DNA for the mutants was then used to transfect HEK293 cells using Fugene (Roche, Indianapolis, IN) and stable cell lines were made for further assays under G418 selection.

Homology Model of rat and mouse TAAR₁

The homology model for rTAAR₁ was created using the Prime software package (Schrödinger Inc.). The model was based on a sequence alignment of rTAAR₁ with the sequence of human β₂AR, for which a crystal structure was recently solved (Protein Data Bank accession code 2RH1), created using the “align GPCR” module (26, 27). The modeling procedure involves side chain rotamer optimization and closure of chain breaks due to gaps in the sequence alignment using a previously published loop building and optimization algorithm (39). After building the complete model, additional side chain rotamer optimization, followed by backbone and side chain energy minimization, was performed on all non-conserved residues. The homology modeling program relies on the OPLS all atom force field (40–42) and a Generalized Born solvent model (43, 44) to evaluate the energy of different conformations and select the lowest energy structure as the final model.

Synthesis

Detailed synthetic procedures and characterization information for novel compounds 8, 9, 11, and 12 are described in the supporting information.

In Vitro cAMP Agonist Activity Assay

Compounds were tested using the Hithunter cAMP XS kit (DiscoveRx) as described previously (24). Data were reported relative to 1 and expressed as %T₁AM. The activity of 1 at 10μM was set as 100 %T₁AM. Concentration-response curves were plotted and EC₅₀ values were calculated with Prism software (GraphPad, San Diego, CA). Standard error of the mean was calculated from the EC₅₀ and E_max values of each independent triplicate experiment by use of Prism Software.