Predicting compounds that interact with the 2 known agonist-induced conformations of the human β1-adrenoceptor

Abstract

The β1-adrenoceptor exists in at least 2 agonist-stabilized conformational ensembles: a “catecholamine” ensemble induced via the intrahelical binding site through which catecholamines and most agonists act and a “secondary” ensemble of conformations through which CGP12177 stimulates agonist responses. Several β-ligands stimulate agonist responses through both conformations, resulting in biphasic concentration responses, but little is known about the structure-activity relationship of such ligands. Using a structure-activity hypothesis built on the predicted poses CGP12177 and 3 biphasic agonists (alprenolol, oxprenolol, and bucindolol), predictions based on ligand similarity and structural compatibility reasoning were made about 11 other β1-ligands not yet tested for secondary conformation interaction and examined in radioligand binding and functional assays using human β1- and β2-adrenoceptors. Although the predictions matched with pharmacology in only 6/11 of cases, 3 novel compounds were found to induce an active-state secondary conformation. A CGP12177 derivative (methyl-pyrrole replacing the cyclic urea motif) retained catecholamine site antagonism with secondary site activation. Carteolol (related to CGP12177) and bunitrolol (similar to alprenolol) activated both conformations with biphasic concentration responses. Bunolol (CGP12177 derivative lacking nitrogen in the bicyclic system), as predicted, was a neutral antagonist with no secondary site activation. Moprolol and some bucindolol analogs appeared as conventional agonists, whereas other alprenolol and bucindolol analogs lost all receptor interaction. In a β1-adrenoceptor mutant (β1-V189T-L195Q-W199Y) where secondary site CGP12177 and pindolol interaction is lost, the 3 novel secondary-site compounds were also no longer able to stimulate secondary conformation responses, suggesting that there is a common TM4 secondary conformation-inducing interaction site.

Significance Statement

The β1-adrenoceptor exists in 2 agonist-stabilized, pharmacologically distinguishable conformations. This study pinpointed the interaction site through which the alternative conformation is stabilized and suggested and evaluated additional ligands, thus providing possible molecular determinants.

1. Introduction

The agonist actions of several β1-adrenoceptor (AR) ligands cannot be explained by interactions at a single site or conformation.¹ Low concentrations of certain β-AR ligands (eg, pindolol² and CGP12177³) block catecholamine responses (via the catecholamine conformation), whereas higher concentrations stimulate partial agonist responses via a secondary site.4, 5, 6 Furthermore, catecholamine conformation full and partial agonist responses are readily inhibited by β-antagonists, whereas secondary conformation agonist responses are relatively resistant to inhibition.^5,7,8 In cat, rodent, and human myocardium, pindolol stimulates biphasic agonist responses.^9,10 Studies with cloned receptors and knockout animals confirmed that all responses are occurring via β1-ARs.11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 Thus, the β1-AR exists in at least 2 agonist-stabilized pharmacological conformations and there are 4 measurable pharmacological phenomena that allow the additional β1-secondary conformation to be distinguished from the conventional β1-catecholamine conformation: (A) in the same system (cells, tissues, or whole animals), secondary conformation agonist responses occur at substantially higher concentrations (ie, with lower affinity) than the concentration required to bind to the catecholamine conformation (ie, EC₅₀ > K_D, where K_D can be measured from radioligand binding or inhibition of agonist responses); (B) inhibition of secondary conformation agonist responses requires substantially higher concentrations of β-antagonist than inhibition of catecholamine-conformation agonist responses; (C) some ligands stimulate biphasic agonist responses where the first component is readily inhibited and the second component relatively resistant to β-antagonism; (D) as to date all secondary conformation agonists bind to the catecholamine conformation with higher affinity (point (A) above), in the presence of a catecholamine site agonist, increasing concentrations of secondary conformation agonist inhibit the catecholamine response at low concentrations and then stimulate their own agonist response at higher concentrations, creating a dip in the curve not reconcilable with a single site of action.

The human β3-AR also has a secondary conformation,²² whereas to date there is no pharmacological evidence of a β2-AR secondary conformation.23, 24, 25

We know several things about β1-AR secondary conformations. First, several ligands stabilize active-state secondary conformations, including alprenolol, bucindolol, carazolol, carvedilol, CGP12177, cyanopindolol, LY362884, oxprenolol, pindolol, SDZ21009, and SR59230A.^{5,7,8,10,11,15,16,19,21,24,26,27} Second, β1-AR secondary conformations exist in animal and human heart and blood vessels (see references above), so although the physiological relevance remains unknown, it is a genuine pharmacological phenomenon. Third, carvedilol therapeutic human plasma concentration (eg, for heart failure and ischemic heart disease) is 100 ng/mL (300 nM²⁸), a concentration at which carvedilol stabilizes a secondary conformation.²⁹ Fourth, although β-blockers inhibit agonist responses at both conformations (albeit at different concentrations) the rank order of affinity differs, confirming the secondary conformation as a distinct entity, not a “mirror” of the catecholamine conformation.²¹ Fifthly, CGP12177 and pindolol analog studies have highlighted certain chemical moieties required for their secondary conformation agonism.^30,31 Sixth, amino acid residues toward the extracellular end of transmembrane 4 (TM4: V189, L195, and W199, that are notably conserved between species for both β1- and β3-AR but are different in β-ARs where no secondary conformation has been described²⁵) are required for both CGP12177 and pindolol secondary conformation stimulation suggesting that both ligands stabilize a common secondary conformation. However, it is unknown whether a plethora of “β1-AR secondary sites” exists, accessed by different ligands, or if there is just one “secondary site” and V189/L195/W199 are important for all secondary conformation agonists.

Little is known about the structure-activity relationship of secondary conformation ligands and even less regarding how ligands access both the catecholamine and secondary sites. One explanation for biphasic responses (eg, pindolol) is that at low concentrations, a molecule occupies the catecholamine site as a typical high-affinity partial agonist. Higher concentrations allow a second molecule to occupy a distinct site (with lower affinity) that directly stabilizes a different overall higher-efficacy receptor conformation. Alternatively, the second molecule could occupy a site that behaves allosterically, augmenting the catecholamine site efficacy (ie, allosterically augmenting its own efficacy, but not affinity). Both explanations hold for CGP12177, where catecholamine site occupation usually results in high-affinity neutral antagonism and secondary site occupancy either causes a direct (lower affinity) stimulation of an active-state conformation of the receptor or an allosteric augmentation of very low (undetected) catecholamine site efficacy.

Given the plethora of structural data for β-ARs, we investigated the nature and existence of secondary sites. Based on a structure-activity hypothesis built on the predicted poses of known secondary ligands, we predicted, then pharmacologically tested, whether other known (but untested) β-AR compounds would stabilize an active-state conformation and then determined whether this was via the same site as the “extra-TM4” site used by CGP12177 and pindolol.

2. Terminology

The observed pharmacological catecholamine and secondary conformations are, as all receptor conformational states, structurally most likely a set of closely related conformations, and will be abbreviated to “CCE” (catecholamine conformation ensemble) and “SCE” (secondary conformation ensemble). Going forward, SCE will be used for discussion of pharmacological phenomena. In contrast, for discussion of the actual site through which the SCE is stimulated, V189/L195/W199 site or known site 2 (“KS2”; see beginning of Results for an explanation of this terminology) will be used. The physical site for ligand-amino acid interactions where, eg, catecholamines bind to the β1-AR, will be termed “intrahelical binding site” (IBS³²) and refers to the general cavity and not specific ligand-amino acid interactions within it, to better distinguish it from the “extra-TM4” V189/L195/W199 site.

3. Materials and methods

3.1. Materials

³H-CGP12177 (NET1061250UC), Steadylite Plus (luciferase detection), Microscint 20, and Ultima Gold XR scintillation fluid were from Revvity (formerly Perkin-Elmer). Cimaterol (0435), CGP20712A (1024), bucindolol (2658), and oxprenolol (3288) were from Tocris. VL01 (Mcule-7804650113, Key Organics P-18694452), VL03 (Mcule-5582827213, MedChemExpress P-588278446), VL04 (Mcule-286344860, MedChemExpress P-588276645), VL11 (Mcule-2229578802, Chembridge P-3794668), VL12 (Mcule-4273841503, Chembridge P-6086407), and VL13 (Mcule-9006078691, Chembridge P-5272537) were obtained via Mcule (Budapest, Hungary). VL05 (Amb17760690), VL06 (Amb24208310), VL07 (Amb6342301), and VL08 (Amb9653) were from Ambinter. VL09 (EN300-1268038) and VL10 (Z31380545) were from Enamine. Figure 1 shows the structures of the compounds used in this study, and Supplemental Tables 1 and 2 list their SMILES and IUPAC names, respectively. CGP12177 (C125), ICI118551 (I127), alprenolol (A8625), pindolol (P0778), and all other reagents were from Sigma Aldrich.

Fig. 1.

Two-dimensional chemical structures of known and predicted compounds used in this study. The VL compound chemical moieties mentioned in the Results and Discussion are highlighted with ellipses in order to make identifying the moieties easier. Ellipse colors match the font colors of the parental compounds’ names.

3.2. Docking calculations

Human β1-AR was modeled using MODELLER³³ for both active and inactive conformations. The crystallographic structure with PDB ID 2VT4 was used as template to model the inactive conformation, whereas PDB ID 3SN6 was used for the active conformation. The Gαs subunit was added to the β1-AR active-state model by superimposing this model with the ternary complex of the β2-AR (PDB ID 3SN6) and removing the receptor portion and small-molecule ligand of 3SN6. Swissdock³⁴ was used to dock CGP12177 to the entire β1-AR protein complex, including the Gαs. As CGP12177 is a hydrophilic ligand and therefore unlikely to pass through the lipid bilayer, we focused on the predicted binding sites accessible from the extracellular space. CGP12177 was then manually docked to the binding pocket between TM3 and TM4 (KS2), using the docking poses from Swissdock as reference. The residues around CGP12177 were afterward minimized using the CHARMM27 forcefield in MOE2019.³⁵

3.3. Similarity searches for analogs

Cartblanche³⁶ was used to search for purchasable compounds with (1) the same substructure or (2) similarity based on the Tanimoto coefficient to known compounds with access to the SCE. The compounds were manually curated for insights into possible structure-activity relationships for the secondary conformation.

3.4. Cell culture

Clonal cell lines of Chinese hamster ovary (CHO)-K1 cells (RIDD: CVCL_0214) stably transfected with the wild-type human β1-AR and a CRE-luciferase reporter gene (CHO-β1-CRE-luc), cells stably transfected with the human β1- or β2-AR and CRE-SPAP reporter gene (CHO-β1-CRE-SPAP, CHO β2-CRE-SPAP), and cells expressing the CRE-SPAP reporter gene but with no transfected receptor (CHO-CRE-SPAP) were used.^19,23 In addition, stable mixed populations of cells expressing human wild type β1-WT (wild type), β1-AR with 3 mutations (β1-V189T-L195Q-W199Y²⁵) were also used. These were generated by transfecting a monolayer of CHO-K1 cells stably expressing a CRE-SPAP reporter in a T75 with 10 ng DNA in 100 μL Lipofectamine and 8 mL OPTIMEM as per manufacturer’s instructions. Cells were selected for 3 weeks using G418 (1 mg/mL for receptor) and hygromycin (200 μg/mL for CRE-SPAP reporter gene) during which time there were passaged at least twice until confluent T75 or T175 flasks of cells were obtained. Selection antibiotics were then removed.

All CHO cells were grown in Dulbecco’s modified Eagle’s medium nutrient mix F12, containing 2 mM L-glutamine and 10% fetal calf serum at 37 °C in humidified 5% CO₂: 95% air atmosphere. With the exception of cells under active selection for the generation for the stable mixed populations, cells were grown in flasks and plates in the absence of antibiotics. Mycoplasma contamination is intermittently monitored within the laboratory (negative), but cell lines were not tested routinely with each experiment.

3.5. ³H-CGP12177 whole-cell binding

The affinity (K_D, concentration required to bind half of the receptors) of ³H-CGP12177 was determined by saturation binding (using 0.008–8.911 nM ³H-CGP12177 and using 10 μM propranolol to determine nonspecific binding). The affinity for competing ligands was determined by incubating the competing ligand in the presence of a fixed concentration of ³H-CGP12177. Briefly, cells were grown to confluence in tissue-culture-treated white-sided 96-well view plates. The media was removed and nonradioligands in 100 μM serum free media (sfm = Dulbecco’s modified Eagle’s medium nutrient mix F12 containing 2 mM L-glutamine) at twice final concentration were added to the wells (including 100 μL 20 μM propranolol to wells used to determine nonspecific binding). This was immediately followed by the addition of 100 μL ³H-CGP12177 (thus a total volume of 200 μL and a 1:2 dilution in the well), as previously described.³⁷ Cells were incubated for 2 hours at 37 °C before being washed with 2 × 200 mL cold (4 °C) PBS. A white bottom was applied to convert the wells to white-bottomed, white-sided wells, 100 μL Microscint 20 was then added to each well, and a clear sealant top was applied. The plates left for at least 8 hours in the dark before being counted on a Topcount for 2 minutes per well. Total binding (6 wells) and nonspecific binding (6 wells, 10 μM propranolol final concentration) was determined in each plate.

3.6. CRE-luciferase production

Cells were plated into tissue-culture-treated white-sided 96-well plates, grown to confluence and CRE-luciferase production determined after 5 hours as previously described.²¹ Briefly, the media was removed and replaced with 100 μL sfm or sfm containing a final concentration of antagonist and incubated for 15 minutes at 37 ˚C. Agonist (in 10 μL sfm, at 10 times final concentration) was then added to the wells (1:10 dilution in the wells), positive control (isoprenaline final well concentration 10 μM), and the plates incubated for 5 hours at 37 °C in a humidified 5% CO₂: 95% air atmosphere. After 5 hours, all well contents were removed, a white bottom applied to the plate, 40 μL luclite reagent (= 20 μL Steadylite Plus + 20 μL PBS containing 1 mM Ca²⁺ and 1 mM Mg²⁺) were added to each well, and the plates kept in the dark for 5 minutes before being read on a Topcount for 2 seconds per well.

3.7. CRE-SPAP production

Cells were plated into tissue-culture-treated clear 96-well plates, grown to confluence and CRE-SPAP production determined as previously described.³⁸ Briefly, at 24 hours, the media was removed and 100 μL sfm added to each well (to serum starve the cells) for a second 24 hours before experimentation. Sfm was removed from wells, and 100 μL sfm or 100 μL antagonist in sfm at final concentration added to the wells and the plate, incubated for 15 minutes at 37 °C. Agonist (in 10 μL, at 10 times final concentration) was then added to the relevant wells (1:10 dilution in the wells), the positive control (isoprenaline, final concentration 10 μM) to control wells, and the plates incubated for 5 hours at 37 °C in a humidified 5% CO₂ : 95% air atmosphere. After 5 hours, all well contents were removed and replaced with 40 μL sfm and incubated for 1 hour (37 °C in a humidified 5% CO₂ : 95% air atmosphere). Plates were then placed in a 65 °C oven for 30 minutes to destroy all endogenous phosphatase, cooled, 100 μL 5 mM pNPP in diethanolamine buffer added to each well, and read on a Dynatech MRX plate reader at 405 nm once the yellow color had started developing.

3.8. Data analysis

3.8.1. Whole-cell binding

The K_D of ³H-CGP12177 was determined by saturation binding fitted using the nonlinear regression program GraphPad Prism 10 (GraphPad) to equation 1:

$$\begin{array}{c}\text{SB}=\end{array}\frac{\left(\mathrm{A}\times {\mathrm{B}}_{\max }\right)}{\left(\mathrm{A}+{\mathrm{K}}_{\mathrm{D}}\right)}$$ (1)

where A is the concentration of ³H-CGP12177, B_max is the maximal specific binding, and K_D is the dissociation constant of ³H-CGP12177.

The affinity of the other ligands was determined from competition binding. A sigmoidal concentration-response curve was then fitted to the data using GraphPad Prism 10, and the IC₅₀ was determined as the concentration required to inhibit 50% of the specific binding using equation 2.

$$\begin{array}{c}\%\text{uninhibited}\text{binding}=\end{array}\begin{array}{ccc}100& -\frac{\left(100\times \left[\mathrm{A}\right]\right)}{\left(\left[\mathrm{A}\right]+{\text{IC}}_{50}\right)}& +\text{NS}\end{array}$$ (2)

where [A] is the concentration of the competing ligand, IC₅₀ is the concentration at which half of the specific binding of ³H-CGP12177 has been inhibited, and NS is the nonspecific binding.

From the IC₅₀ value and the known concentration of ³H-CGP12177, a K_D value (concentration at which half the receptors are bound by the competing ligand) was calculated using equation 3 (Cheng-Prusoff equation):

$$\begin{array}{cc}{\mathrm{K}}_{\mathrm{D}}& =\end{array}\frac{{\text{IC}}_{50}}{1+\left(\left[{}^3\mathrm{H}-\text{CGP}12177\right]/{\mathrm{K}}_{\mathrm{D}}\left({}^3\mathrm{H}-\text{CGP}12177\right)\right)}$$ (3)

3.8.2. CRE-luciferase and CRE-SPAP functional experiments:

Many agonist responses were best described by a one-site sigmoidal agonist concentration-response curve. Curves were fitted to the data using equation 4:

$$\begin{array}{cc}\text{Response}=& \frac{{\text{E}}_{\text{max}}\times \left[\mathrm{A}\right]}{{\text{EC}}_{50}+\left[\mathrm{A}\right]}\end{array}$$ (4)

where E_max is the maximal response, [A] is the agonist concentration, and EC₅₀ is the concentration of agonist that produces 50% of the maximal response.

Antagonist K_D values were then calculated from the parallel shift of the agonist concentration responses in the presence of a fixed concentration of antagonist using equation 5 (Gaddum equation):

$$\begin{array}{cc}\text{DR}=& 1+\frac{\left[\mathrm{B}\right]}{{\text{K}}_{\text{D}}}\end{array}$$ (5)

where DR (dose ratio) is the ratio of the agonist concentration required to stimulate an identical response in the presence and absence of a fixed concentration of antagonist [B].

In experiments where 3 different fixed concentrations of the same antagonist were used, Schild plots were constructed using equation (6):

$$\begin{array}{cc}\log \left(\text{DR}-1\right)=& \log \left[\mathrm{B}\right]-\log \left({\mathrm{K}}_{\mathrm{D}}\right)\end{array}$$ (6)

These points were then fitted to a straight line. A slope of 1 then indicates competitive antagonism.³⁹

Where clear partial agonism was present (eg, Fig. 3D), the partial agonist affinity was calculated by the method of Stephenson⁴⁰ using equation 7.

Fig. 2.

(A) Top view of the β1-AR (green ribbons; TMs numbered with roman numerals) and the 3 extracellular binding sites identified via Swissdock (S1–S3). Docked poses of CGP12177 are shown with purple carbons. (B) Side view of the β1-AR (green ribbons) toward TM4, showing the residues (sticks with green carbons) surrounding the most plausible pose of CGP12177 (purple carbons) in KS2.

$$\begin{array}{ccc}{\mathrm{K}}_{\mathrm{D}}\text{partial}\text{agonist}& =& \frac{\mathrm{Y}\times \left[\mathrm{P}\right]}{1-\mathrm{Y}}\\ \text{where}\mathrm{Y}& =& \frac{\left[{\mathrm{A}}_2\right]-\left[{\mathrm{A}}_1\right]}{\left[{\mathrm{A}}_3\right]}\\ & & \end{array}$$ (7)

where [P] is the concentration of the partial agonist, [A₁] is the concentration of the agonist at the point where the fixed partial agonist causes the same response, [A₂] is the concentration of agonist causing a given response above that achieved by the partial agonist, and [A₃] is the concentration of the agonist, in the presence of the partial agonist, causing the same stimulation as [A₂].

Some agonist’s responses were clearly best described by a 2-site stimulatory concentration response (eg, Fig. 11B); thus equation 8 was fitted to these data:

$$\begin{array}{ccc}\%\text{maximal}\text{stimulation}& =& \frac{\left[\mathrm{A}\right]\times \mathrm{N}}{(\left[\mathrm{A}\right]+\text{EC}{1}_{50})}+\frac{\left[\mathrm{A}\right]\times (100-\mathrm{N})}{(\left[\mathrm{A}\right]+\text{EC}{2}_{50})}\end{array}$$ (8)

Fig. 11.

CRE-SPAP production in (A, B) CHO-β1-CRE-SPAP cells and (C, D) CHO-β2-CRE-SPAP cells in response to (A) and (C) cimaterol in the absence and presence of 3, 10, 30, 100, and 300 nM VL04, and (B, D) VL04 in the absence and presence of (B) 100 nM CGP20712A and (D) 10 nM, 100 nM, and 1 μM ICI118551. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 11, (C) 6, and (D) 6 separate experiments. The Schild slope for (D) was 1.11 ± 0.07 (n = 3).

where N is the percentage of site 1, [A] is the concentration of agonist, and EC1₅₀ and EC2₅₀ are the respective EC₅₀ values for the 2 agonist sites.

At other times, the concentration-response curves (eg, Fig. 3E) clearly contained 2 components: an inhibitory response followed by a stimulatory response; thus, a 2-site analysis was performed using equation 9:

$$\text{Response}=\text{Basal}+\left(\text{control}-\text{Basal}\right)·\left[1-\frac{\left[\mathrm{A}\right]}{\left(\left[\mathrm{A}\right]+{\text{IC}}_{50}\right)}\right]+{\mathrm{S}}_{\mathrm{MAX}}·\left[\frac{\left[\mathrm{A}\right]}{\left(\left[\mathrm{A}\right]+{\text{EC}}_{50}\right)}\right]$$ (9)

where "Basal" is the response in the absence of agonist, "control" is the response to a fixed concentration of control (eg, cimaterol 30 nM), [A] is the concentration of agonist, IC₅₀ is the concentration of agonist that inhibits 50% of the response to control, EC₅₀ is the concentration of agonist that caused a half maximal stimulation, and S_MAX is the maximum stimulation of this component.

Fig. 3.

(A) Inhibition of ³H-CGP12177 whole-cell binding by CGP12177 in CHO-β1-CRE-SPAP cells. The concentration of ³H-CGP12177 is 0.71 nM. (B–F) CRE-luciferase production in CHO-β1-CRE-luciferase cells in response to (B) cimaterol in the absence and presence of 10 nM CGP20712A; (C) CGP12177 in the absence and presence of 1 μM CGP20712A; (D) cimaterol in the absence and presence of 3, 10, or 30 nM CGP12177; (E) CGP12177 in the absence and presence of 10, 30, or 100 nM cimaterol; and (F) bucindolol in the absence and presence of 1μM CGP20712A. Bars represent basal CRE-luciferase production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 10, (C) 10, (D) 6, (E) 6, and (F) 7 separate experiments.

4. Results

4.1. Molecular modeling to identify ligands with secondary-site interaction

After docking CGP12177 to the entire surface of the β1-AR, we observed 3 clusters of poses on the receptor. The classical IBS³² (labeled S1 in Fig. 2A), a second site was located between helices III and IV (S2 in Fig. 2A), and a third one between helices I and VII (S3 in Fig.2A). In the latter site, none of the poses of CGP12177 formed favorable interactions, and we therefore continued with the second site, which corresponds to KS2.³² The residues surrounding CGP12177 in KS2 are shown in Fig. 2B, and mutations of these residues are likely to affect the binding of CGP12177 to the secondary pocket. These residues are F129, E132, L133, S136, F191, I194, L195, M196, W199, H198, and R200. Of these, only L133 (TM3) and L195 and W199 (TM4) differ between the human β1- and β1-AR receptor. This is in keeping with earlier studies where L195 and W199 were found to be essential for CGP12177’s secondary conformation interaction. As modeling has suggested only this one plausible extracellular non-IBS interaction, this suggests that there is indeed likely only one non-IBS or secondary interaction site accessed by CGP12177 on the β1-AR, and it is KS2.

Based on 5 already known compounds that activate an SCE at the β1-AR (alprenolol, bucindolol, CGP12177, oxprenolol, and pindolol), we searched for analogs and retrieved 3 analogs of CGP12177, 1 analog of pindolol, 4 analogs of bucindolol, and 4 analogs of alprenolol and oxprenolol from the ZINC20 library.

The ligands were sourced and passed to the pharmacology team who analyzed the compounds blind to the predictions made by molecular modeling.

4.2. ³H-CGP1277 whole-cell binding in stable cell lines

The K_D value for ³H-CGP12177 and receptor expression levels have previously been determined in the stable cell lines as 0.15 nM (79 fmol/mg protein) in CHO-β1-CRE-luc cells, 0.42 nM (1146 fmol/mg protein) in CHO-β1-CRE-SPAP cells, and 0.17 nM (466 fmol/mg protein) in CHO-β2-CRE-SPAP cells.^19,23 The binding affinity for other ligands (including non-radiolabelled CGP12177) was determined from competition binding (Table 1, Fig. 3). The high selectivity of the β1 antagonist CGP20712A for β1-AR, and the β2 antagonist ICI118551 for β2-AR, confirms the presence of the β1- and β2-AR in the CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP lines, respectively. The concentration of ³H-CGP12177 used in these experiments will only detect ligands binding at the high-affinity catecholamine confirmation of the human β1-AR; thus, Table 1 shows the affinity of these ligands for the catecholamine conformation of the β1-AR and does not give any information about secondary-site interaction.

Table 1.

Affinity (log K_D values) of β-AR ligands for the β1 and β2-AR obtained from ³H-CGP12177 whole-cell binding to CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP cells, respectively

The values are mean ± s.e.mean for n separate experiments. The β2 over β1-selectivity is also given; thus, ICI118551 has 603-fold higher affinity for the β2- than β1-adrenoceptor, whereas CGP20712A (at 0.0008-fold) has 1202-fold higher affinity for the β1-adrenoceptor.

Ligand	Log KDβ1	n	Log KDβ2	n	β2-Selectivity
CGP20712A	−8.80 ± 0.07	6	−5.72 ± 0.05	6	0.0008
ICI118551	−6.74 ± 0.06	8	−9.52 ± 0.06	6	603
Cimaterol	−6.43 ± 0.05	6	−7.11 ± 0.07	6	4.8
CGP12177	−9.29 ± 0.07	7	−9.75 ± 0.07	7	2.9
Alprenolol	−8.20 ± 0.06	6	−9.46 ± 0.05	6	18.2
Bucindolol	−9.24 ± 0.07	6	−10.29 ± 0.07	6	11.2
Oxprenolol	−8.06 ± 0.06	6	−9.37 ± 0.09	6	20.4
Pindolol	−8.71 ± 0.07	6	−9.52 ± 0.09	6	6.5
VL01	−9.45 ± 0.08	7	−10.53 ± 0.07	7	12.0
VL03	−8.29 ± 0.04	7	−10.09 ± 0.04	7	63.1
VL04	−8.64 ± 0.09	7	−9.76 ± 0.10	7	13.2
VL05	−7.62 ± 0.06	7	−7.91 ± 0.06	7	1.9
VL06	−5.95 ± 0.06	7	−6.01 ± 0.08	7	1.1
VL07	−8.21 ± 0.07	7	−9.19 ± 0.05	6	9.5
VL08	IC50>−4	5	−5.49 ± 0.16	5
VL09	−8.51 ± 0.04	7	−8.96 ± 0.08	7	2.8
VL10	−6.73 ± 0.05	7	−7.05 ± 0.06	6	2.1
VL11	IC50>−4	5	−5.05 ± 0.14	5
VL12	IC50>−4	5	−5.17 ± 0.16	5
VL13	No binding to 100 μM	5	No binding to 100 μM	5

4.3. CHO-β1-CRE-luciferase responses

4.3.1. Pharmacological characterization of the CCE and SCEs with cimaterol and CGP12177

Cimaterol (log EC₅₀ −8.07 ± 0.03, 106.2 ± 2.0% isoprenaline maximum, n = 19, Table 2) stimulated an agonist response in the CHO-β1-CRE-luciferase cells that was inhibited by CGP20712A with high affinity (log K_D CGP210712A −9.47 ± 0.12, n = 12; Fig. 3), and by ICI118551 with lower affinity (log K_D ICI118551 −7.33 ± 0.06, n = 5, Table 3) confirming the cimaterol response was indeed occurring via the β1-AR. CGP12177 also inhibited the cimaterol response as a partial agonist to yield a log K_D value of −9.87 ± 0.06 n = 11 (as calculated by partial agonist method⁴⁰; Fig. 3). Cimaterol is an efficacious agonist, with an affinity (log K_D of −6.43 or 372 nM) and log EC₅₀ of −8.07 (or 8.5 nM), and thus it only needs to bind a few receptors to induce a full agonist response.

Table 2.

Agonist actions of compounds in CHO-β1-CRE-luciferase cells

Log EC₅₀ values are given, with % of the response they generated compared to 10 μM isoprenaline. The values are mean ± s.e.mean for n separate experiments.

Ligand	Log EC50	% isop	n	Log KD CGP20712A	n
Alprenolol	−6.49 ± 0.11	3.2 ± 0.8	8
Bucindolola	EC501 −9.27 ± 0.14 EC502 −7.24 ± 0.11 23.3 ± 2.4% at EC501	23.7 ± 1.7	7	−8.06 ± 0.11	8
CGP12177	−7.54 ± 0.04	54.2 ± 3.5	25	−7.37 ± 0.09	17
Cimaterol	−8.07 ± 0.03	106.2 ± 2.0	19	−9.47 ± 0.12	12
Oxprenolol	−6.28 ± 0.11	5.3 ± 0.8	8
Pindolol	−5.64 ± 0.12	11.7 ± 1.4	8
VL01	−6.78 ± 0.05	19.7 ± 2.4	16	−7.29 ± 0.08	13
VL03	No response		5
VL04	−6.12 ± 0.06	23.7 ± 3.1	13	−7.39 ± 0.06	22
VL05	−7.19 ± 0.08	9.7 ± 0.8	18	−9.54 ± 0.11	15
VL06	−5.85 ± 0.31	2.7 ± 0.8	8
VL07	No response		5
VL08	No response		5
VL09	−6.89 ± 0.03	23.9 ± 2.2	17	−7.84 ± 0.06	18
VL10	−6.81 ± 0.06	13.8 ± 1.5	20	−9.73 ± 0.10	17
VL11	No response		5
VL12	No response		5
VL13	No response		5

^aBucindolol stimulated a biphasic concentration-response curve (Fig. 3F). The log EC₅₀ at both components are given, along with % occurring via EC₅₀1 and % of isoprenaline of the whole response. The K_D value for CGP20712A is for inhibition of the first component.

Table 3.

Log K_D values of ligands determined from a rightward shift of agonist responses to cimaterol or CGP12177 in CHO-β1-CRE-luciferase cells

Values are calculated using the method of Gaddum (eq. 5). The values are mean ± s.e.mean for n separate experiments.

The ratio of affinities for the 2 conformations is also shown. Compounds all have higher affinity for the catecholamine-induced than secondary conformation, eg, VL01 has log 3.55 (ie, 3500-fold) high affinity for the catecholamine rather than secondary conformation, whereas VL10 has log 1.64 (44-fold) higher affinity for the catecholamine site. As in²¹, this corroborates that the secondary conformation is a distinct entity (not a low-affinity “mirror” of the catecholamine conformation), with compounds interacting with it in a distinct manner with both distinct affinity (measured here) and efficacy (previous table).

	Catecholamine Conformation		Secondary Conformation		Log ratio of affinities
	Cimaterol as agonist		CGP12177 as agonist
	Log KD	n	Log KD	n
CGP12177	−9.87 ± 0.06a	11
CGP20712A	−9.47 ± 0.12	12	−7.37 ± 0.09	17	2.10
ICI118551	−7.33 ± 0.06	5	−5.76 ± 0.05	6	1.57
VL01	−10.19 ± 0.05	12	−6.64 ± 0.07a	7	3.55
VL03	−8.91 ± 0.05	6	−6.38 ± 0.01	6	2.53
VL04	−9.12 ± 0.09	8	−5.70 ± 0.10∗	8	3.42
VL05	−8.00 ± 0.04a	8	No shift 10 μM	5	>3
VL06	−6.46 ± 0.04	5	No shift 10 μM	5	>1.46
VL07	−8.87 ± 0.06	6	−6.16 ± 0.04	9	2.71
VL08	<−5	5	No shift 10 μM	5
VL09	−8.84 ± 0.03a	6	−6.03 ± 0.10a	7	2.81
VL10	−7.21 ± 0.03a	7	−5.57 ± 0.10a	7	1.64
VL11	<−5	6	No shift 10 μM	5
VL12	No shift at 10 μM	5	No shift 10 μM	5
VL13	No shift at 10 μM	6	No shift 10 μM	5

^aLigand with partial agonism and therefore the K_D was calculated using the method of Stephenson (eq. 7).

CGP12177 stimulated a partial agonist response (log EC₅₀ −7.54 ± 0.04, 54.2 ± 3.5% isoprenaline maximum, n = 25; Fig. 3, Table 2) that required higher concentrations of CGP20712A and ICI118551 to inhibit and thus yielded log K_D values of −7.37 ± 0.09 (n = 17) and −5.76 ± 0.05 (n = 6), respectively (Table 3). These values are very similar to previous data.²¹ Thus, within this cell line, the affinity of CGP12177 as determined from inhibition of an agonist (log K_D −9.87 or 0.13 nM) or as determined from binding (0.42 nM from saturation binding and log K_D −9.29 or 0.51 nM, from competition binding) is at odds with its EC₅₀ (log EC₅₀ −7.54 or 29 nM). For a partial agonist activating a single conformation the K_D and EC₅₀ should be the same. However, here when half the receptors are bound (0.13 nM), there is no stimulatory response, yet 29 nM stimulates half the maximum response obtainable with CGP12177, suggesting that the high-affinity binding of CGP12177 and the stimulatory response are acting at 2 different sites of the receptor (evidence (A) for SCE activation as outlined in the Introduction). Furthermore, the concentrations of antagonists (CGP20712A and ICI118551) required to inhibit the CGP12177 agonist response were far greater than those required to inhibit the cimaterol response (yielding K_D values 126 and 37-fold higher respectively), again suggesting that cimaterol and CGP12177 agonist responses are occurring at different sites or conformations of the receptor (evidence (B) for SCE activation as outlined in the Introduction). Thus, this demonstrates the 2 pharmacologically observable conformation ensembles of the β1-AR: cimaterol is acting through the catecholamine conformation (the CCE) where agonist responses are readily inhibited by antagonists (including CGP12177), whereas the CGP12177 agonist response is occurring through a secondary conformation (an active-state SCE) where agonist responses are relatively resistant to inhibition. Finally, if increasing concentrations of CGP12177 are examined in the presence of a fixed concentration of cimaterol (Fig. 3E), it can be seen that low concentrations of CGP12177 inhibit the cimaterol response (due to high-affinity antagonism of the catecholamine conformation) whereas higher concentrations of CGP12177 stimulate an agonist response, as first shown by,¹¹ evidence (D) for SCE activation as outlined in the Introduction. Bucindolol had a biphasic concentration response (Fig. 3F, evidence (C) for SCE activation as outlined in the Introduction).

The pharmacological response to other compounds fell into 4 distinct groups: VL03/07, VL05/06/10, VL01/04/09, and VL08/11/12/13 (compound structures are shown in Fig. 1).

4.3.2. Determination of the site of action of VL03 and VL07

VL03 and VL07 inhibited cimaterol-induced responses in CHO-β1-CRE-luciferase cells with high affinity (Fig. 4B; Table 3). They both also inhibited the CGP12177-induced response but required higher concentrations to do so (Fig. 4C); thus again, these ligands had higher affinity for the catecholamine conformation ensemble (Table 3) of the β1-AR. VL03 and VL07 did not stimulate any agonist response themselves in CHO-β1-CRE-luciferase cells, although as expected from their affinity, low concentrations of VL03 and VL07 inhibited the cimaterol response as seen in Fig. 4D. There is therefore no evidence (A–D/Introduction) for SCE activation.

Fig. 4.

(A) Inhibition of ³H-CGP12177 whole-cell binding by VL03 in CHO-β1-CRE-SPAP cells. The concentration of ³H-CGP12177 is 0.71 nM. (B–E) CRE-luciferase production in CHO-β1-CRE-luciferase cells in response to (B) cimaterol in the absence and presence of 10 nM and100 nM VL03; (C) CGP12177 in the absence and presence of 1 μM and 10 μM VL03; and (D) VL03 in the absence and presence of 10, 30, or 100 nM cimaterol. Bars represent basal CRE-luciferase production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 5, (C) 5, and (D) 5 separate experiments.

4.3.3. Determination of the site of action of VL10, VL05, and VL06

VL10 (Fig. 5B), VL05, and VL06 all inhibited cimaterol CHO-β1-CRE-luciferase responses, thus acting as antagonists of the catecholamine-stabilized conformation ensemble (Table 3). Even with the maximum concentration of ligand possible (10 μM), this has little impact on the CGP12177 agonist response (Fig. 5C). Although the response to VL06 was too small to accurately quantify, the agonist responses to VL10 and VL05 were readily antagonized by CGP20712A to yield K_D values of −9.73 and −9.54, respectively (Fig. 5D; Table 2). VL10, VL05, and VL06 inhibited a fixed concentration of cimaterol at concentrations similar to those required to stimulate their agonist response. There is therefore no evidence (A–D) for SCE activation, and all the evidence suggests that β1-AR agonist interactions of VL10, VL05, and VL06 are occurring through the same conformation as cimaterol, ie, the CCE.

Fig. 5.

(A) Inhibition of ³H-CGP12177 whole-cell binding by VL10 in CHO-β1-CRE-SPAP cells. The concentration of ³H-CGP12177 is 0.53 nM. (B–E) CRE-luciferase production in CHO-β1-CRE-luciferase cells in response to (B) cimaterol in the absence and presence of 10 μM VL10; (C), CGP12177 in the absence and presence of 10 μM VL10; (D) VL10 in the absence and presence of 1, 3, or 10 nM CGP20712A; and (E) VL10 in the absence and presence of 10, 30, or 100 nM cimaterol. Bars represent basal CRE-luciferase production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 7, (C) 7, (D) 9, and (E) 7 separate experiments.

4.3.4. Determination of the site of action of VL01, VL04, and VL09

VL01 (Fig. 6B), VL04, and VL09 inhibited cimaterol β1-CRE-luciferase responses with high affinity (log K_Ds −10.19, −9.12, and −8.84, respectively; Table 3). This is similar to the values obtained from ³H-CGP12177 whole-cell binding above, showing these compounds bind to the catecholamine-stabilized conformation ensemble with high affinity. VL01 (Fig. 6C), VL04 and VL09 are also able to inhibit the CGP12177 β1-CRE-luciferase responses but with lower affinity (−6.64, −5.70, and −6.03, respectively) and are therefore able to bind to the SCE, but as with other antagonists, the affinity for this secondary conformation is less. From the high concentrations of VL01, VL04, and VL09 required to inhibit the CGP12177 response (Fig. 6C), it is also evident that VL01, VL04, and VL09 have agonist properties of their own.

Fig. 6.

(A) Inhibition of ³H-CGP12177 whole-cell binding by VL01 in CHO-β1-CRE-SPAP cells. The concentration of ³H-CGP12177 is 0.71 nM. (B–E) CRE-luciferase production in CHO-β1-CRE-luciferase cells in response to (B) cimaterol in the absence and presence of 1 and 3 nM VL01; (C) CGP12177 in the absence and presence of 1 and 10 μM VL01; (D) VL01 in the absence and presence of 1 μM CGP20712A; and (E) VL01 in the absence and presence of 10, 30, or 100 nM cimaterol. Bars represent basal CRE-luciferase production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 6, (C) 5, (D) 9, and (E) 6 separate experiments.

VL01 (Fig. 6D), VL04, and VL09 all stimulated partial agonist responses in CHO-β1-luciferase cells (Table 2). Considerably higher concentrations of ligand were required to stimulate this response than inhibit the catecholamine conformation (evidence (A) for SCE activation). The concentration of antagonist (CGP20712A) required to inhibit these partial agonist responses was also high (Table 2; evidence (B) for SCE activation). When all 3 were examined in the presence of a fixed concentration of cimaterol, low concentrations inhibited the agonist actions of cimaterol, whereas higher concentrations were needed to stimulate the agonist response (Fig. 6E; evidence (D) for SCE activation). Thus VL01, VL04, and VL09 were appearing pharmacologically similar to CGP12177, ie, having agonist actions via an SCE.

4.3.5. Determination of the site of action of VL08, VL11, VL12, and VL13

The affinity of these compounds as determined by inhibition of ³H-CGP12177 was too low to be determined (Table 1). Although 10 μM VL08 and VL11 was just about able to cause a rightward shift of the cimaterol response (Fig. 7), it was too small to quantify. VL12 and VL13 (10 μM) were not able to inhibit the response at all (Fig. 7). None of the compounds were able to inhibit the CGP12177 response (up to 10 μM, ie, the maximum concentration possible, Fig. 7). Likewise, they stimulated no agonist response, and minimal or no inhibition of the fixed cimaterol response was seen (Fig. 7). These ligands therefore barely interacted with the β1-AR at the CCE, and there is no evidence (A–D) of any SCE activation.

Fig. 7.

(A, E) Inhibition of ³H-CGP12177 whole-cell binding by (A) VL08 and (E) VL13 in CHO-β1-CRE-SPAP cells. The concentration of ³H-CGP12177 is (A) 0.62 nM and (E) 0.71 nM. (B–D, F–H) CRE-luciferase production in CHO-β1-CRE-luciferase cells in response to (B) cimaterol in the absence and presence of 10 μM VL08; (C) CGP12177 in the absence and presence of 10 μM VL08; (D) VL08 in the absence and presence of 10 nM cimaterol; (F) cimaterol in the absence and presence of 10 μM VL13; (C) CGP12177 in the absence and presence of 10 μM VL13: (D) VL13 in the absence and presence of 10 nM cimaterol. Bars represent basal CRE-luciferase production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 5, (B) 5, (C) 5, (D) 5, (E) 5, (F) 6, (G) 5, and (H) 5 separate experiments.

4.4. CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP responses

Ligands were evaluated in a separate second system (CRE-SPAP cells) for several reasons. First, this would confirm the β1-AR responses to ligands in a separate system, and show that the existence of the secondary site is not a product of receptor expression levels. Second, it allowed comparison between the β1-AR and β2-AR responses in the same cell background. Third, as a very different reporter product is being measured, it would demonstrate that the reporter protein itself is not interfering with the observed pharmacology. Finally, the CHO-β1-CRE-SPAP cell line has a high (supraphysiological) receptor expression level and so allows for the detection of small partial agonist responses, be it those occurring via the CCE or the SCE for the β1-AR. Responses in different cell lines with high and low receptor expression levels allow for further evidence (A–D) for SCE activation to be gathered.

Cimaterol stimulated a response in both CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP cells (Table 4) that was readily inhibited by CGP20712A in the CHO-β1 cells and by ICI118551 in the CHO-β2 cells, in keeping with previous studies, and thus demonstrating the presence of the respective receptors in each cell line (Table 5). In the CHO-β1-SPAP cells, the CGP12177 response (log EC₅₀ −8.71 [Table 4] greater than log K_D −9.29 [Table 1], ie, evidence (A) for SCE activation) was (as expected with higher receptor expression levels) proportionately greater than in β1-luciferase cells (54% vs 74% isoprenaline) and left shifted (more potent; log EC₅₀ −7.54 and −8.71) but once again required higher concentrations of antagonists to cause a rightward shift (yielding poorer K_D values, evidence (B) for SCE activation), demonstrating the presence of the SCE of the β1-AR. In the CHO-β2-cells, CGP12177 also stimulated a partial agonist CRE-SPAP response; however, this was readily inhibited by antagonists yielding similar log K_D values when measured as a shift of the cimaterol or CGP12177 response, demonstrating a lack of the CGP12177-induced SCE at the β2-AR (see ratios in Table 5). Alprenolol, oxprenolol, and pindolol, as previously reported,^19,24 stimulated β1-biphasic responses (Table 4; evidence (C) for SCE activation).

Table 4.

Agonist actions of compounds in CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP cells

Log EC₅₀ values are given, with % of the response compared with 10 μM isoprenaline. Where obtained, the log K_D values for inhibition of agonist response by CGP20712 (β1) or ICI118551 (β2) are also given. The values are mean ± s.e.mean for n separate experiments.

	CHO-β1-CRE-SPAP					CHO-β2-CRE-SPAP Cells
	Log EC50	% Isoprenaline	n	Log KD CGP20712A	n	Log EC50	% Isoprenaline	n	Log KD ICI118551	n
Alprenolola	EC501 −8.34 ± 0.07 EC502 −6.44 ± 0.14 62.5 ± 2.7% at EC501	56.7 ± 3.7	8	−8.44 ± 0.11	7	−9.58 ± 0.06	49.9 ± 2.2	6	−9.68 ± 0.03	5
Bucindolol	−9.12 ± 0.04	84.5 ± 2.4	9	−7.68 ± 0.09	9	−9.70 ± 0.04	70.3 ± 2.0	8	−8.81 ± 0.03	6
CGP12177	−8.71 ± 0.08	74.0 ± 3.7	17	−7.02 ± 0.12	12	−9.57 ± 0.08	53.0 ± 2.4	19	−9.56 ± 0.06	10
Cimaterol	−9.26 ± 0.05	82.8 ± 3.1	20	−8.96 ± 0.12	17	−10.14 ± 0.05	95.3 ± 2.1	16	−9.76 ± 0.03	11
Oxprenolola	EC501 −8.41 ± 0.03 EC502 −5.68 ± 0.11 79.8 ± 2.4% at EC501	67.8 ± 3.9	7	−8.77 ± 0.07	7	−9.42 ± 0.05	40.1 ± 2.3	8	−9.92 ± 0.08	6
Pindolola	EC501 −8.75 ± 0.05 EC502 −5.91 ± 0.11 62.7 ± 2.4% at EC501	81.0 ± 2.6	11	−8.59 ± 0.11	7	−9.73 ± 0.07	61.8 ± 2.8	7	−9.87 ± 0.08	6
VL01	−8.56 ± 0.10	69.6 ± 2.5	14	−7.59 ± 0.09	20	−9.85 ± 0.08	52.1 ± 3.5	14	−9.13 ± 0.05	23
VL03	No response		5			No response		5
VL04a	EC501 −9.06 ± 0.13 EC502 −6.39 ± 0.14 52.1 ± 2.8% at EC501	75.2 ± 3.6	18	−8.59 ± 0.11	12	−9.59 ± 0.13	55.1 ± 2.7	11	−9.53 ± 0.04	13
VL05	−8.30 ± 0.06	62.3 ± 3.8	13	−9.06 ± 0.15	12	−8.42 ± 0.09	63.2 ± 3.2	13	−9.78 ± 0.10	17
VL06	−6.47 ± 0.10	53.6 ± 3.1	15	−9.11 ± 0.08	8	−6.44 ± 0.15	15.2 ± 1.2	12
VL07	−6.55 ± 0.17	26.5 ± 1.9	12			−9.30 ± 0.14	13.1 ± 2.0	10
VL08	No response		5			No response		6
VL09a	EC501 −9.15 ± 0.08 EC502 −6.61 ± 0.21 69.4 ± 3.2% at EC501	80.1 ± 4.1	13	−8.50 ± 0.11	12	−9.35 ± 0.13	65.8 ± 4.0	11	−9.84 ± 0.07	14
VL10	−7.73 ± 0.07	66.9 ± 2.5	19	−8.88 ± 0.12	15	−7.40 ± 0.06	28.7 ± 2.0	13	−9.93 ± 0.09	9
VL11	No response		5			No response		6
VL12	No response		5			No response		5
VL13	No response		6			No response		4

^aAlprenolol, oxprenolol, pindolol, VL04, and VL09 stimulated biphasic concentration-response curves in the CHO-β1-CRE-SPAP cells (Fig. 11). The log EC₅₀ at both components are given, along with % occurring via EC₅₀1 and % of isoprenaline of the whole response. The K_D value for CGP20712A is for inhibition of the first component.

Table 5.

Log K_D values of ligands determined from a rightward shift of agonist responses to cimaterol or CGP12177 in CHO-β1-CRE-SPAP and CHO-β2-CRE-SPAP cells

Values are calculated using the method of Gaddum (eq. 5). The values are mean ± s.e.mean for n separate experiments.

The ratio of affinities for the 2 conformations is also shown (as in Table 3). Whereas this ratio shows that ligand affinity for the 2 conformations varies between 65 and 316-fold in the CHO-β1 cells, the differences in antagonist affinity when cimaterol and CGP12177 are agonists in the CHO-β2 cells are only 1.6 to 3.8-fold in keeping with a single conformation.

	CHO-β1-CRE-SPAP					CHO-β2-CRE-SPAP
	Log KD Catecholamine conformation Cimaterol as Agonist	n	Log KD Secondary Conformation CGP12177 as Agonist	n	Log Ratio of Affinities	Log KD (Cimaterol as Agonist)	n	Log KD (CGP12177 as Agonist)	n	Log Ratio of Affinities
CGP20712A	−8.96 ± 0.12	17	−7.02 ± 0.12	12	1.94	−6.05 ± 0.08	6	−5.74 ± 0.14	6	0.31
ICI118551	−7.30 ± 0.07	10	−5.49 ± 0.10	6	1.81	−9.76 ± 0.03	11	−9.56 ± 0.06	10	0.20
VL01	−10.17 ± 0.14a	14	b			−11.08 ± 0.13a	16	b
VL03	−8.75 ± 0.05	11	−6.25 ± 0.06	10	2.50	−10.33 ± 0.04	21	−10.12 ± 0.05	30	0.21
VL04	−9.15 ± 0.11a	13	b			−9.94 ± 0.07a	16	b
VL05	b		b			b		b
VL06	b		b			−6.34 ± 0.17a	8	−5.93 ± 0.15	4	0.41
VL07	−8.59 ± 0.05	17	−6.19 ± 0.11	8	2.40	−9.47 ± 0.06	17	−8.89 ± 0.07	14	0.58
VL08	No shift 10 μM	5	No shift 10 μM	5		−5.68 ± 0.14	7	No shift 10 μM	5
VL09	−9.71 ± 0.20a	12	b			−9.36 ± 0.10a	12	b
VL10	−6.93 ± 0.25	7	b			−7.43 ± 0.09	7	−6.86 ± 0.10	10	0.57
VL11	No shift 10 μM	6	No shift 10 μM	6		No shift 10 μM	5	No shift 10 μM	5
VL12	No shift 10 μM	5	No shift 10 μM	5		−5.43 ± 0.26	5	No shift 10 μM	5
VL13	No shift 10 μM	4	No shift 10 μM	5		No shift 10 μM	5	No shift 10 μM	4

^awhere the ligand had partial agonism and therefore the K_D was calculated using the method of Stephenson (eq. 7).

^bThe partial agonist stimulation was too great to enable a calculation of K_D value.

4.5. VL03 and VL07

VL03 did not stimulate an agonist response in either cell line (Table 4); however, it inhibited the cimaterol responses in both CHO-β1 and CHO-β2 cells with high affinity, as expected from its inhibition of ³H-CGP12177 (Fig. 8; Tables 1 and 5). VL03 also inhibited the CGP12177 agonist responses. The β2-response was readily inhibited to yield a similar K_D value for ICI118551 very similar to when cimaterol was the agonist, suggesting both cimaterol and CGP12177 are acting through the same conformation, whereas the CHO-β1 CGP12177 response required a higher concentration of VL03 to cause a rightward shift, yielding a poorer K_D value, in keeping with lower affinity inhibition of the SCE as for most β-antagonists (Fig. 8; Table 5). There is therefore no evidence (A–D) for any SCE activation. VL07 stimulated small agonist responses in both cell lines (Table 4) as previously reported,¹⁹ and comparison of its EC₅₀ and K_D would suggest that its β1-AR agonist response is occurring via an SCE (evidence (A) for SCE activation).

Fig. 8.

CRE-SPAP production in (A) and (B) CHO-β1-CRE-SPAP cells and (C) and (D) CHO-β2-CRE-SPAP cells in response to (A) and (C) cimaterol in the absence and presence of 1 nM, 10 nM, 100 nM, and 1 μM VL03, and (B) and (D) CGP12177 in the absence and presence of 1 nM, 10 nM, 100 nM, and 10 μM VL03. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 8, (C) 7, and (D) 10 separate experiments. The Schild slopes were (A) 0.96 ± 0.05 (n = 7), (C) 0.99 ± 0.04 (n = 7), and (D) 1.00 ± 0.03 (n = 7).

4.6. VL10, VL05, and VL06

As expected, the partial agonist response to VL10 was one again substantially greater in the high expressing β1-SPAP cells (67% isoprenaline maximum vs 14% in the CHO-β1-CRE-luciferase cells; Fig. 9; Table 4), but K_D values for inhibition of cimaterol and CGP12177 were just possible to determine. The agonist response to VL10 response was however readily inhibited by CGP20712A. The agonist responses to VL05 and VL06 were likewise proportionally greater, and therefore allowing inhibition by CGP20712A to be evaluated. VL10, VL05, and VL06 were all readily inhibited by CGP20712A, suggesting that these agonist responses were occurring via the CCE with no evidence (A–D) for any SCE activation.

Fig. 9.

CRE-SPAP production in (A) and (B) CHO-β1-CRE-SPAP cells and (C–E) CHO-β2-CRE-SPAP cells in response to (A) and (C) cimaterol in the absence and presence of 10 μM VL10 and (D) CGP12177 in the absence and presence of 10 μM VL10 and (B) and (E) VL10 in the absence and presence (B) 10 nM CGP20712A and (E) 10 nM ICI118551. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 7, (B) 13, (C) 7, (D) 10, and (E) 8 separate experiments.

4.7. VL01, VL04, and VL09

VL01 was able to inhibit the β1-cimaterol response to yield a K_D value similar to that measured by ³H-CGP12177 whole-cell binding (Fig. 10), suggesting high-affinity binding to the catecholamine conformation. However, its log EC₅₀ − 8.56 (Table 4) was substantially higher than its K_D measurements (log K_D − 9.45 [Table 1] and log K_D − 10.17 [Table 5], evidence (A) for SCE activation). Its partial agonism was similar to that of CGP12177, rendering a measurement of affinity at the SCE unobtainable. In the CHO-β1-CRE-SPAP cells, VL01 responses were more resistant to inhibition (evidence (B) for SCE activation), yielding K_D values for CGP20712A similar to those when CGP12177 was the agonist (Fig. 10; Tables 4 and 5). In the CHO-β2-CRE-SPAP cells, VL01-stimulated β2-agonist responses were readily inhibited by ICI118551 to yield a similar value to those obtained when cimaterol and CGP12177 were agonists. This suggests that all ligands are acting at the same β2-conformation (Fig. 10; Tables 4 and 5).

Fig. 10.

CRE-SPAP production in (A, B) CHO-β1-CRE-SPAP cells and (C, D) CHO-β2-CRE-SPAP cells in response to (A, C) cimaterol in the absence and presence of 1, 3, and 10 nM VL01, and (B, D) VL01 in the absence and presence of (B) 100 nM, 1 μM, and 10 μM CGP20712A and (D) 1, 10, and 100 nM ICI118551. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 10, (B) 11, (C) 9, and (D) 11 separate experiments. The Schild slopes were (B) 0.94 ± 0.17 (n = 4) and (D) 0.96 ± 0.07 (n = 4).

VL04 and VL09 inhibited cimaterol responses in both the β1 and β2 cell lines, but again their increased partial agonist response meant that measurements of VL04 and VL09 antagonist affinity in the presence of CGP12177 were not possible. In the CHO-β2 cells, VL04 and VL09 agonist responses were readily inhibited by ICI118551, yielding similar K_D values to all others obtained, suggesting all compounds were interacting at the same conformation. In the CHO-β1-CRE-SPAP cells, both of these compounds had a biphasic concentration-response curve (Fig. 11, evidence (C) for SCE activation). The first component of this was more readily inhibited by CGP20712A than the secondary component, in a manner very similar to that seen with pindolol.¹⁹

4.8. β1-WT and β1V189T-L195Q-W199Y mutation data

A previous study has suggested that certain amino acids located at the extracellular end of TM4 (V189, L195, and W199) are important for the β1-SCE interaction of CGP12177 and pindolol.²⁵ These are the same residues that also emerged from the unbiased docking calculations reported above. When these were mutated to those of the β2-AR (namely T, Q, and Y respectively), CGP12177 was no longer able to activate secondary site response and appeared to interact with the β1-AR as a conventional partial agonist at the same site as cimaterol. The biphasic partial agonist responses to pindolol also became monophasic suggesting that binding to the secondary site was no longer possible. Thus, in the β1-V189T-L195Q-W199Y receptor, all evidence (A–D) for SCE activation was lost.²⁵ VL01, VL04, and VL09 responses were therefore compared in stable mixed populations of cells expressing the β1-WT or the triple mutant β1-V189T-L195Q-W199Y receptor.

4.9. ³H-CGP12177 whole-cell binding in β1-WT and β1-V189T-L195Q-W199Y cells

³H-CGP12177 saturation binding in 6 stable mixed populations expressing the β1-AR and 6 expressing the β1-V189T-L195Q-W199Y receptors yielded K_D values for ³H-CGP12177 of 0.17 ± 0.01 nM (n = 6) and 0.27 ± 0.02 nM (n = 6), respectively (Fig. 12). From competition binding, the K_D values obtained for cimaterol, CGP20712A, VL01, VL04, and VL09 were similar with both receptors (Table 6). However, as expected, ICI118551 had higher affinity for the β1-V189T-L195Q-W199Y mutant than β1-WT (Table 6). As previously noted, the presence of the V189T mutation (either alone or in combination with other mutations)²⁵ increases the affinity of ICI118551, thus distinguishing these populations of cells from those expressing the β1-WT receptor.

Fig. 12.

(A) and (D) ³H-CGP12177 saturation binding in stable mixed population of CHO cells expressing (A) β1-AR and (D) β1-V189T-L195Q-W199Y mutant receptor. (B, C, E, and F) CRE-SPAP production in stable mixed population of CHO cells expressing (B, C) β1-AR and (E, F) β1-V189T-L195Q-W199Y mutant receptor, (B) and (E) in response to cimaterol in the absence and presence 10 nM CGP20712A and 3 nM CGP1277 and (C) and (F) CGP12177 in the absence and presence of (C) 1 μM CGP210712A and (F) 10 nM CGP20712A. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 6, (B) 6, (C) 9, (D) 6, (E) 6, and (F) 9 separate experiments.

Table 6.

Affinity of β-adrenoceptor ligands for the β1-WT and β1-V189T-L195Q-W199Y receptors expressed in stable mixed populations of CHO cells obtained from ³H-CGP12177 whole-cell binding (left half) and from rightward shift of a cimaterol CRE-SPAP response (right half)

Values are mean ± s.e.mean for n separate experiments. From binding studies, for affinity of ³H-CGP12177 (K_D) was determined from saturation binding and was 0.170 ± 0.007 nM (log −9.77, n = 6) for β1-WT and 0.272 ± 0.020 nM (log −9.57, n = 6) for β1-V189T-L195Q-W199Y. A higher affinity for ICI118551 in β1-V189T-L195Q-W199Y compared with β1-WT was expected as all V189T mutant receptors have this property.²⁵

	3H-CGP12177 whole cell binding β1-WT		3H-CGP12177 whole cell binding β1-V189T-L195Q-W199Y		CRE-SPAP production β1-WT		CRE-SPAP production β1-V189T-L195Q-W199Y
	Log KD	n	Log KD	n	Log KD	n	Log KD	n
CGP12177	(−9.77)		(−9.57)		−10.11 ± 0.07a	6	−10.28 ± 0.11a	6
CGP20712A	−9.40 ± 0.03	6	−9.50 ± 0.06	6	−9.44 ± 0.07	6	−9.80 ± 0.12	6
Cimaterol	−6.63 ± 0.06	7	−6.76 ± 0.05	7
ICI118551	−6.95 ± 0.07	7	−7.60 ± 0.02	7	−7.04 ± 0.11	6	−8.14 ± 0.12	6
VL01	−9.85 ± 0.03	8	−9.56 ± 0.03	8	−10.24 ± 0.10a	9	−10.09 ± 0.06a	10
VL04	−8.93 ± 0.03	10	−8.76 ± 0.03	10	−9.30 ± 0.09a	10	−9.56 ± 0.06a	11
VL09	−8.70 ± 0.04	9	−8.61 ± 0.02	9	−9.17 ± 0.11a	10	−9.27 ± 0.07a	11

^awhere the ligand had partial agonism and therefore the K_D was calculated using the method of Stephenson (eq. 7).

4.10. CRE-SPAP responses in β1-WT and β1-V189T-L195Q-W199Y cells

At both receptors, the cimaterol responses were inhibited by both CGP20712A and CGP12177 with high affinity, yielding log K_D value similar to those obtained from ³H-CGP12177 whole-cell binding (Fig. 12; Tables 6 and 7) and similar to those obtained in the stable cell lines. As expected, the affinity of ICI118551 was increased in the triple mutant compared with β1-WT. In the β1-WT cells, the CGP12177 response again required higher concentrations of antagonist to inhibit the response (evidence (B) for SCE activation), and the CGP12177 response (log EC₅₀ −8.30) was itself right-shifted compared with the K_D (log K_D − 9.77; evidence (A) for SCE activation), again confirming the presence of the SCE in these stable mixed populations of cells. In the β1-V189T-L195Q-W199Y cells, however, the partial agonist CGP12177 response (log EC₅₀ –9.67) was similar to the K_D value in these cells (log K_D − 9.57 measured from ³H-CGP12177 binding above = loss of evidence (A) compared with β1-WT). Furthermore, the CGP12177 response is more readily inhibited by CGP20712A, yielding K_D values for CGP20712A similar to that obtained when cimaterol was the agonist (= loss of evidence (B) compared with β1-WT; Fig. 12; Table 7). Overall, this suggests that at the β1-V189T-L195Q-W199Y receptor, CGP12177 is stimulating its agonist response via the same conformation as cimaterol (as seen in the β2-AR). Thus, in the β1-V189T-L195Q-W199Y receptor, the SCE appears absent, while leaving the CCE pharmacologically intact.

Table 7.

Agonist actions of compounds obtained from β1-WT and β1-V189T-L195Q-W199Y receptors expressed in stable mixed populations of CHO cells

Log EC₅₀ values are given, with % of the response compared to 10 μM isoprenaline. Log K_D values for inhibition of the agonist response by CGP20712A are also given. The values are mean ± s.e.mean for n separate experiments.

	β1-WT					β1-V189T-L195Q-W199Y
	Log EC50	% isoprenaline	n	Log KD CGP20712A	n	Log EC50	% isoprenaline	n	Log KD CGP20712A	n
Cimaterol	−8.25 ± 0.04	82.8 ± 1.5	8	−9.44 ± 0.07	6	−8.33 ± 0.05	79.2 ± 2.1	9	−9.80 ± 0.12	6
CGP21177	−8.30 ± 0.06	45.2 ± 2.8	9	−7.37 ± 0.09	13	−9.67 ± 0.06	27.9 ± 1.7	9	−9.30 ± 0.10	13
VL01a	EC501 −9.65 ± 0.11 Log EC502 −6.70 ± 0.12 49.0 ± 2.3 % at EC501	42.9 ± 3.0	15	−8.42 ± 0.14	17	−9.59 ± 0.06	30.4 ± 1.3	16	−8.84 ± 0.13	20
VL04a	EC501 −9.10 ± 0.14 EC502 −6.22 ± 0.14 41.9 ± 2.6 % at EC501	39.5 ± 3.0	13	−8.79 ± 0.18	11	−9.18 ± 0.09	26.7 ± 1.6	16	−9.18 ± 0.14	12
VL09a	EC501 −9.29 ± 0.15 EC502 −6.76 ± 0.14 53.9 ± 3.3% at EC501	45.2 ± 4.7	10	−8.79 ± 0.13	11	−9.20 ± 0.09	35.6 ± 1.8	9	−9.43 ± 0.16	11

^aVL01, VL04, and VL09 stimulated a biphasic concentration-response curve in the β-WT cells (Fig. 13). The log EC₅₀ at both components are given, along with along with % occurring via EC₅₀1 and % of isoprenaline of the whole response. The K_D value for CGP20712A is for inhibition of the first component.

When the responses to VL04 and VL09 were examined in these cells (Fig. 13; Table 7), a biphasic response was seen for both ligands in the β1-WT stable mixed populations of cells, with the first component being more readily inhibited than the second, just as in the stable CHO-β1-CRE-SPAP cell line (evidence (C) for SCE activation). In the β1-V189T-L195Q-W199Y cells, the response was a monophasic partial agonist response (= loss of evidence (C) compared with β1-WT) that was readily inhibited by CGP20712A; thus, the more antagonist-resistant secondary component appeared to be missing (Fig. 13; Table 7, = loss of evidence (B) for SCE activation). For VL01, in contrast to the response in the stable cell lines, the response in the β1-WT stable mixed population appeared biphasic (evidence (C) for SCE activation), although the antagonist-resistant secondary component also appeared to have disappeared in the β1-V189T-L195Q-W199Y cells (= loss of evidence (C) compared with β1-WT.

Fig. 13.

CRE-SPAP production in stable mixed population of CHO cells expressing (A–C) β1-AR and (D–F) β1-V189T-L195Q-W199Y mutant receptor in response to (A) and (D) VL01 in the absence and presence 100 nM CGP20712A, (B) and (E) VL04 in the absence and presence 100 nM CGP210712A and (C) and (F) VL09 in the absence and presence of 100 nM CGP20712A. Bars represent basal CRE-SPAP production and that in response to 10 μM isoprenaline and other compounds alone as listed with each graph. Data points are mean ± SD of triplicate determinations and are representative of (A) 12, (B) 11, (C) 11, (D) 12, (E) 12, and (F) 11 separate experiments.

4.11. Lack of responses in CHO-CRE-SPAP cells

In CHO-CRE-SPAP cells (ie, cells expressing the reporter gene but no transfected receptor), whereas forskolin stimulated a response that was 3.0 ± 0.1-fold over basal (n = 5), there were no responses to any of the VL compounds (n = 5 each ligand). As previously reported, there are no responses to cimaterol, CGP12177, ICI118551, or CGP20712A in these cells, either. This confirms that all responses seen were occurring via the transfected receptors.

5. Discussion

The β2-AR appears as a conventional receptor. CGP12177 affinity (³H-CGP12177 binding log K_D −9.77 from saturation and −9.75 from competition binding) is similar to its partial agonist log EC₅₀ value (−9.57). Responses are similarly inhibited (ICI118551 log K_D of −9.76 when cimaterol, and −9.56 when CGP12177 is the agonist) suggesting interaction at a single receptor conformation. VL03 was a high-affinity neutral antagonist. VL07, VL06, VL10, VL04, VL01, VL09, and VL05 had increasing degrees of partial agonism (with responses readily inhibited by ICI118551), whereas VL08, VL11, VL12, and VL13 barely interacted with the receptor (all compounds depicted in Fig. 1). Thus, VL compounds behaved as conventional β2-ligands interacting at a single conformation—the CCE/IBS.

For β1-ARs, several pharmacological anomalies reveal an agonist-stabilized SCE. First, CGP12177’s high affinity (eg, log K_D −9.87 from inhibition of cimaterol in β1-CRE-luciferase cells) differs from its partial agonist log EC₅₀ value (−7.54), suggesting CGP12177 agonist activation via a different conformation (an SCE, evidence (A)). Second, CGP12177 agonist responses are relatively resistant to inhibition (required 126-fold higher CGP20712A concentrations to inhibit CGP12177 than cimaterol responses), suggesting CGP12177 stimulation is not via the CCE (evidence (B)). Third, low CGP12177 concentrations inhibit cimaterol but high CGP21177 concentrations stimulate a response (Fig. 3E, evidence (D)), which cannot be reconciled with single conformation interaction. Thus, 2 β1-AR conformations are demonstrated: a high-affinity CCE where agonist responses are readily antagonized, and a low-affinity CGP12177-stabilized SCE, where agonist responses are more resistant to antagonism.¹

Although several ligands stimulate β1-AR SCE responses, the structure-activity relationship (SAR) surrounding SCE interaction is relatively unknown. CGP12177 (most studied), pindolol, bucindolol, oxprenolol, and alprenolol, which all activate an SCE, were chosen as starting points to correlate the presence of certain functional groups with β1-AR SCE activation. To provide molecular explanations for chemical moiety importance, we determined potential SCE-inducing sites for the ligands from modeling and found only one likely site—similar to the G protein-coupled receptor (GPCR) allosteric pocket KS2,³² which also overlaps with the pharmacological β1-TM4-V189T-L195Q-W199Y site.²⁵ From the predicted CGP12177 binding mode, we suggested how certain functional groups interact with KS2 residues, conducted similarity searches for additional β-AR ligands, and evaluated them for potential compatibility with KS2. After predicting whether the compounds would/would not bind to KS2 and thus induce an SCE, we pharmacologically evaluated them. Table 8 and Fig. 14 contain a summary of the predictions and results, and we provide a more in-depth reasoning and explanation in the following.

Table 8.

Summary of compound-β1-AR predictions, actual pharmacological outcomes, evidence for SCE activation (A–D) as outlined in the Introduction) from data presented in this study from either CHO-β1-luciferase or CHO-β1-SPAP cells, and whether the prediction of SCE activation and pharmacological outcome of SCE activation were correct

Compound	Parent Compound	Prediction of Active-State SCE Stabilization	Pharmacological Outcome			Prediction vs Pharmacological Outcome for SCE Activation
			CCE/IBS conformation	SCE conformation	Evidence of SCE activation
Alprenolol			Agonist (biphasic)	Agonist (biphasic)	C
Bucindolol			Agonist (biphasic)	Agonist (biphasic)	C
CGP12177			High-affinity neutral antagonist	Agonist	ABD
Cimaterol			Agonist	a	None
Oxprenolol			Agonist (biphasic)	Agonist (biphasic)	C
Pindolol			Agonist (biphasic)	Agonist (biphasic)	C
CGP20712A			High-affinity neutral antagonist	Low-affinity neutral antagonist	None
VL01	CGP12177	Yes	High-affinity neutral antagonist	Agonist	ABCD	Correct
VL03 (bunolol)	CGP12177	No	High-affinity neutral antagonist	Low-affinity neutral antagonist	None	Correct
VL04 (carteolol)	CGP12177	Yes	Agonist (biphasic)	Agonist (biphasic)	ABCD	Correct
VL05 (moprolol)	Oxprenolol	--	Agonist	a	None	n/a
VL06	Bucindolol	Yes	Agonist	a	None	Not correct
VL07 (propranolol)		No	High-affinity neutral antagonist	Very weak activation	A	Not correct
VL08	Alprenolol	No	No receptor interaction	No receptor interaction	None	Correct
VL09 (bunitrolol)	Alprenolol	No	Agonist (biphasic)	Agonist (biphasic)	ABCD	Not correct
VL10	Bucindolol	Yes	Agonist	a	None	Not correct
VL11	Bucindolol	Yes	No receptor interaction	No receptor interaction	None	Not correct
VL12	Alprenolol	No	No receptor interaction	No receptor interaction	None	Correct
VL13	Alprenolol	No	No receptor interaction	No receptor interaction	None	Correct

n/a: prediction not made, this compound was found during similarity searches but its SCE outcome was not predicted ahead of pharmacological studies.

^aAlthough it is not possible to absolutely exclude SCE activation (it may be obscured by significant CCE/IBS activation), the affinity values of CGP20712A suggest inhibition of agonist action at the CCE/IBS only.

Fig. 14.

Summary of the findings: the protein shown is PDB ID 7BVQ, and the IBS has been colored in cyan, and was defined as any residue within 5 Å around the ligand in the structure, carazolol. Residues in magenta are the ones that emerged from our docking calculations, and correspond to KS2. Compound classification is according to the pharmacological results reported in this work.

VL03 (bunolol), CGP12177 related, was not predicted to stimulate a β1-AR SCE. It lacks bicyclic system nitrogen atoms, rendering it unable to provide hydrogen bond donor functionalities for receptor interactions at KS2, and we would thus have assumed no binding to KS2 and therefore no activation of an SCE. However, VL03 was a conventional β-antagonist—a high-affinity competitive neutral CCE antagonist and a low-affinity competitive neutral antagonist for CGP12177’s SCE. VL07 (propranolol), predicted not to stimulate a β1-AR SCE for the same reason, was a high-affinity CCE neutral antagonist and a low-affinity SCE antagonist in the low-receptor-expressing CHO-β1-CRE-luciferase cells (with similar K_D values to previous studies²¹). In the high-receptor-expressing cells (CHO-β1-CRE-SPAP), where weak agonist responses are more readily detected, VL07 stimulated a weak partial agonist response. VL07 (pharmacologically evaluated blind to its identity) yielded K_D and EC₅₀ values almost identical to previous studies with propranolol.^19,21,37 Also, as previously determined,¹⁹ the agonist response (log EC₅₀ −6.55) differed from its affinity (log K_D −8.21 from binding studies and −8.59 from inhibition of cimaterol at the CCE), suggesting that this very weak stimulation is occurring via a β1-AR SCE (evidence (A)), in contrast to our prediction.

VL10, bucindolol related, has a fluorine rather than the nitrile group. Our prediction suggested the nitrile was not crucial for KS2 binding, and so we would still observe induction of an active-state SCE by VL10. VL10 had lower affinity than bucindolol (−6.73 vs −9.24) and interacted with both the β1-AR CCE and CGP12177-induced-SCE (inhibiting both cimaterol and CGP12177 responses), although as with all compounds, with lower SCE affinity. VL10 stimulated agonist responses that, as expected, were more potent in the higher expressing CHO-β1-CRE-SPAP cells (log EC₅₀ values −6.81 and −7.73 in the low and high-receptor-expressing cells, respectively), but these responses were readily inhibited by CGP20712A. Furthermore, Fig. 5E suggests VL10 and cimaterol competition at the same site. Altogether, this suggests that VL10’s agonist actions occur via the β1-AR CCE. Although an SCE activation at higher VL10 concentrations cannot be absolutely excluded (it could be obscured by the large CCE agonist response), there is no evidence for SCE activation (compare Figs 3E and 6E with Fig. 5E; VL10 agonist responses readily inhibited by CGP20712A; agonist response same or left shifted of K_D). Thus, contrary to our prediction, the nitrile appears important for bucindolol’s SCE activation.

For VL06, bucindolol related, with a methoxy group instead of the nitrile, this substitution again reduced affinity and VL06’s agonist response was readily inhibited by CGP20712A. Thus, VL06, like VL10, is a conventional CCE partial agonist. VL06 and VL10 indicate that the indole is not important for SCE activation but suggest a role for the nitrile group. VL11, bucindolol related, lacking both the nitrile (replaced by a 1-hydroxy-ethoxy group) and the central β-hydroxy ether, lost all β1-AR interaction, including SCE activation.

VL05 (moprolol) is oxprenolol related but lacking the terminal alkene moiety. VL05 agonist responses were monophasic and readily inhibited by CGP20712A, suggesting conventional CCE partial agonism. Thus, oxprenolol’s alkene tail appears important for its stabilization of an active-state SCE.

VL01, CGP12177 related, has a double bond replacing one side of the cyclic urea, preserving the second nitrogen as an aromatic amine and hydrogen bond donor (ie, indole instead of 1,3-dihydro-2H-benzimidazol-2-one). A previous study suggested the carbonyl group had little pharmacological effect at β1-ARs, but the NH group in position 1 (analogous to indole) was essential for active-state SCE induction.³¹ VL01 was predicted to induce an active-state SCE. VL01 retained β1-AR high affinity (³H-CGP12177 binding log K_D −9.45, −10.19, and −10.17 from cimaterol response inhibition), suggesting CCE high affinity. VL01 was a β1-AR partial agonist, but the log EC₅₀ values (−6.78 (20% isoprenaline) in the low-receptor-expressing cells and −8.56 (70% isoprenaline) in the high-receptor-expressing cells) are considerably different from its CCE high-affinity binding (evidence (A)). VL01 agonist responses were resistant to CGP20712A inhibition (evidence (B)), and low VL01 concentrations inhibited cimaterol, distinct from the high concentration required for agonist action (evidence (D), Fig. 7E). VL01 remained similar to CGP12177—a high-affinity β1-AR CCE antagonist with partial agonist activation via an SCE.

VL04 (carteolol), CGP12177 related, with the urea replaced by an amide positioned to allow receptor interaction, was predicted to stabilize an active-state β1-AR SCE. VL09 (bunitrolol), alprenolol related with a nitrile substituting the ortho-propenyl moiety, was predicted not to induce an active-state SCE. Both compounds were high-affinity CCE antagonists, but with right shifted (evidence (A)), antagonist-resistant (evidence (B)) agonist responses in CHO-β1-CRE-luciferase cells, suggesting both ligands induce an active-state SCE. In the higher-receptor-expressing CHO-β1-CRE-SPAP cells, VL04 and VL09 stimulated biphasic responses (evidence (C)). The first component was readily inhibited by CGP20712A, whereas the second component was resistant to antagonism (Fig. 11B), very similar to findings with alprenolol and pindolol.¹⁹ Thus, both VL04 and VL09 activate an SCE.

VL13 (alprenolol lacking the ortho-propylene substituent), VL08 (alprenolol lacking the ortho-propylene, but with 3 methyl groups in ortho and para positions), and VL12 (alprenolol without the β-hydroxy and a longer aliphatic tail terminating in a methyl-propyl ether instead of the isopropyl group), barely interacted with the β1-AR. We predicted these compounds would not lead to an active-state SCE, but they actually lacked affinity or stimulation for either the CCE or SCE. The β-hydroxy and ortho-propylene groups clearly have important roles, and substituents in the para position (even a small methyl) are detrimental.

CGP12177 and pindolol stabilize a shared β1-AR active-state SCE via the extracellular end of TM4 (V189-L195-W199), and these residues are within the KS2 site. When mutated to the equivalent β2-AR residues (threonine, glutamine, and tyrosine, respectively), the CGP12177 and pindolol active-state SCE is lost (= loss of evidence (A–D) compared with β1-WT).²⁵ As 3 novel compounds were identified that induce a β1-AR active-state SCE, the responses to VL01, VL04, and VL09 were therefore examined in the β1-V189T-L195Q-M199Y mutant receptor to determine if they were activating the receptor’s SCE via the same site, KS2.

The CCE affinity of ligands measured by ³H-CGP12177 binding and inhibition of cimaterol responses were very similar for the wild-type β1-AR (β1-WT) and β1-V189T-L195Q-W199Y receptor, with the exception of ICI118551, where mutation of V189 to threonine increases the ICI118551 affinity.²⁵ Cimaterol responses remained similar in β1-WT and β1-V189T-L195Q-W199Y receptors confirming an unchanged CCE. CGP12177 responses however were more potent (log EC₅₀ −9.67 similar to log K_D −9.57 = loss of evidence (A)), with lower stimulation (28% rather than 45% of isoprenaline) and readily antagonized by CGP20712A at β1-V189T-L195Q-W199Y (= loss of evidence (B)). Thus, CGP12177 appeared as a conventional CCE/IBS partial agonist losing its SCE induction. VL04 and VL09’s biphasic β1-WT responses became monophasic at β1-V189T-L195Q-W199Y (loss of evidence (C)) with EC₅₀ values similar to K_D values, lower overall responses (compared with isoprenaline), and responses readily inhibited by CGP20712A. VL01 had a biphasic response in the β1-WT stable mixed cell populations (Fig. 13A, evidence (C)), similar to CGP12177 in some stable mixed populations (eg, Figure 5a of²⁵), and suggests that in an amplified system, stabilization of both conformations can be measured. Like CGP12177, VL01 responses became high potency, monophasic, lower % isoprenaline, and more readily inhibited in the triple mutant (loss of evidence (C)), suggesting a single active-state CCE, and loss of SCE in β1-V189T-L195Q-W199Y mutant receptor. Thus, VL01, VL04, and VL09 all were unable to induce an active-state SCE in β1-V189T-L195Q-W199Y, and their SCE is induced via the same site as by CGP12177 and pindolol—the TM4 site corresponding to KS2.

Thus, 3 more ligands have been identified that stimulate an active-state SCE of the human β1-AR, VL01, VL04, and VL09. Modeling identified only one potential SCE-inducing site and mutagenesis confirmed that VL01, VL04, and VL09 use the same β1-TM4-V189T-L195Q-W199Y site as CGP12177 and pindolol. This site corresponds to KS2 from earlier work analyzing a large set of receptors,³² which suggested that more class A GPCRs possess a similar cavity. It remains unknown if compounds could be generated that interact with other GPCR KS2 sites and whether SCE inhibition or activation is physiologically relevant.

Conflict of interest

The authors declare no conflicts of interest.