Binding mechanism of nine N-phenylpiperazine derivatives and α_1A-adrenoceptor using site-directed molecular docking and high performance affinity chromatography

Abstract

N-Phenylpiperazine derivatives are widely used as clinical drugs for fighting diseases related to the cardiovascular system by mediating the signal pathway of α₁-adrenoceptor. The binding mechanism of nine N-phenylpiperazine derivatives to α_1A-adrenoceptor was explored using molecular docking and high performance affinity chromatography. The methodology involved homology modelling of the three dimensional structure of α_1A-adrenoceptor, predication of the binding behaviors using LIBDOCK and investigation of the thermodynamic behaviors of the binding by frontal analysis. Molecular docking results showed that Asp¹⁰⁶, Gln¹⁷⁷, Ser¹⁸⁸, Ser¹⁹² and Phe¹⁹³ of the receptor were the main binding sites for the nine N-phenylpiperazine derivatives binding to α_1A-adrenoceptor. The binding was driven by formation of hydrogen bonds and electrostatic forces. The affinity of these derivatives for the receptor depended on the functional groups of an ionizable piperazine, hydrogen bond acceptor and hydrophobic moiety in the ligand structures. Frontal analysis indicated that the association constants of these compounds for the receptor were determined by their structural deviations in the above-mentioned functional groups. Thermodynamic studies presented negative enthalpy and Gibbs free energy changes with a positive entropy change, providing proof that the binding of the derivatives to α_1A-adrenoceptor was mainly driven by electrostatic forces. This result was in line with the binding mechanism predicted by molecular docking. It is possible to explore the binding mechanism of drug candidates specifically binding to α_1A-adrenoceptor using receptor chromatography.

---

Introduction

The adrenergic receptors (ARs) belong to the family of G-protein coupled seven-transmembrane receptors, which serve as the preferred targets for more than fifty percent of drugs approved by the U.S. Food and Drug Administration.^1–3 The receptors are divided into three subclasses: α₁, α₂ and β, and further into several subtypes, such as α_1A, α_1B, α_1D, α_2A, α_2B, α_2C, β₁, β₂, and β₃.^4–7 Among these subtypes, α₁-ARs play the roles of contracting vascular smooth muscle and human prostate smooth muscle, increasing blood pressure, dilating pupils and also regulating cerebral microcirculation.⁸ Regarding these physiological functions of α₁-ARs, ligands, especially antagonists of the receptors, have been introduced to pharmacotherapy and have become current first-line medications with considerable success in curing hypertension.⁹ In this context, the search for new antagonists of α₁-ARs has attracted great attention in medicinal and analytical chemistry.

A series of techniques have been successfully developed for the pursuit of new ligands binding to known targets. These techniques include computer-aided drug discovery and development,¹⁰ high throughput screening assays¹¹ and fragment-based drug discovery.¹² Another validated strategy for searching for new ligands is based on the structure–activity relationship determined by ligand–target interaction analysis. This relationship is further employed to design new candidates with higher affinity and stronger activity towards the same target. Modern technologies, videlicet surface plasmon resonance,¹³ microdialysis¹⁴ and fluorescence methods,¹⁵ have been utilized to explore ligand–target interactions. Affinity chromatography has also proved to be an extremely powerful tool for the same purpose due to the bio-specificity incorporated into the design of the affinity stationary phases and the high specificity, sensitivity and speed of operation resulting from high performance liquid chromatography. Among the affinity chromatographic studies, immobilized human serum albumin is the most widely used stationary phase for revealing the binding of drugs to the protein.¹⁶ More specific stationary phases containing nicotinic acetylcholine receptors,¹⁷ P-glycoprotein¹⁸ and cannabinoid (CB1/CB2) receptor¹⁹ are also constructed for the analysis of ligand–protein interactions by attaching the proteins on the surface of the solid matrix through physical adsorption or covalent bonding.

Our group has found that the stationary phase containing immobilized α₁-ARs has potential in measuring the association constants of drug–receptor interactions.^20,21 Despite the capacity for drug–receptor interaction analysis, the feasibility of the immobilized α₁-ARs in realizing the binding mechanism of drugs to the receptors needs further investigation.

The antagonists of α₁-AR are associated with N-aryl and N-heteroaryl piperazine derivatives.²² Prazosin, terazosin and doxazosin, the highly selective α₁-AR antagonists, are designed by using a piperazine-1,4-diyl moiety. The three compounds are the successful cases of improved therapeutic efficacy as a result of subtype selectivity. The structure–affinity–activity studies of α₁-AR antagonists derived from N-phenylpiperazine compounds are continuously necessary in biochemistry, medicine and biology. This work aimed to simulate the interaction between nine N-phenylpiperazine derivatives and α_1A-AR using site-directed molecular docking. Further work was performed to confirm the validated application of immobilized α_1A-AR in analyzing the binding mechanism of these compounds to the receptor by frontal analysis.

Experimental

Materials and instruments

ANTI-FLAG® M1 agarose affinity gel was purchased from Sigma-Aldrich Co. LLC (Saint Louis, MO, USA). Macroporous silica gel (SPS 300-7, pore size 300 Å, particle size 7.0 μm) was from Fuji Silysia Chemical Company Limited (Tokyo, Japan). N-Phenylpiperazine derivatives (compounds 1–2, 6–9) were synthesized and identified using the method in our previous report.²³ The purities of these compounds were determined to be more than 98% using high performance liquid chromatography. Standards of doxazosin (compound 3), prazosin (compound 4) and terazosin (compound 5) were purchased from the Institute of Drug and Biological Product Control of China (Beijing, China). All other reagents were analytically pure unless stated otherwise.

The chromatographic system consisted of an Agilent 1100 series of apparatus including a binary pump, a column oven, a diode array detector (Waldbronn, Germany) and a Chemistation 5.2 software installation for data acquisition and processing. The ZZXT-A type packing instrument was from Dalian Elite Analytical Instruments Co., Ltd (Dalian, China).

Purification and immobilization of α_1A-AR

Human embryonic kidney 293 (HEK293) cells stably expressing α_1A-AR were prepared and cultured using the method in a previous report.²⁴ The cells were harvested by centrifugation for 10 min with a speed of 3000 rpm at 4 °C. Subsequent treatment was performed by adding three volumes of lysis buffer (50 mM Tris–HCl, 150 mM NaCl, 2 mM DTT, 10% glycerol, pH 7.2) supplemented with a protease inhibitor cocktail (Sigma) to one volume of the obtained cell pellet. The lysate was ruptured in a bead mill for 15 min at 4 °C. Following an additional centrifugation for 20 min at a speed of 30 000 rpm at 4 °C, 50 mL of the cell extract was suspended in 5 mM CaCl₂ and loaded onto a 5 mL anti-FLAG M1 agarose column. The unbound proteins were removed by eluting the column with loading buffer (20 mM Tris, 500 mM NaCl, 10% glycerol) in the presence of 5 mM CaCl₂. Captured proteins were eluted with the loading buffer supplemented with 10 mM EDTA instead of CaCl₂. The obtained protein was confirmed to be α_1A-AR from its molecular weight of 66.0 kDa, determined using sodium dodecyl sulphate polyacrylamide gel electrophoresis.

The purified α_1A-AR was immobilized on the surface of silica gel using a widely reported method.^25,26 Briefly, γ-aminopropyl triethoxysilane was utilized to convert the hydroxyl groups on the gel to amino groups, which were further activated with N,N′-carbonyldiimidazole. The activated gel was suspended in 10.0 mL phosphate buffer (pH = 7.0) in the presence of 4.0 mL α_1A-AR for an extra 2.0 h of reaction. After rinsing with 150 mL phosphate buffer (50 mM, pH 7.0) containing 2.0 M NaCl, the gel was transferred into a 30 mL 1% glycine ethylester solution to remove the residuals of unreacted imidazole groups. The immobilized α_1A-AR was packed into a stainless steel column (50 × 4.6 mm, 7.0 μM) using Tris–HCl buffer (5 mM, pH 7.2) as a slurry and propulsive agent under a pressure of 4.0 × 10⁷ Pa.

Molecular docking

The crystal structure of α_1A-AR was constructed using homology modelling. In this case, the crystal structure of human β₂-AR (PDB entry: 3SN6) at 2.4 Å resolution²⁷ was utilized as a template through the search of the related protein structure using the BLAST (Basic Local Alignment Search Tool) program²⁸ in the Protein Data Bank database. The amino acidic sequence of human α_1A-AR was retrieved from the Swiss-Prot database²⁹ (accession number P35348, entry name ADA1A_HUMAN) and aligned with human β₂-AR. The three dimensional homology modeling was performed using the crystal structural coordinate of the template on the basis of alignment of the target and template sequence of β₂-AR according to the guidelines of Discovery Studio 2.5 (DS 2.5, Accelrys software Inc., San Diego, CA). Subsequent structural evaluation was performed using PROCHECK³⁰ (a program to check the stereochemical quality of protein structures) and PROFILE 3D³¹ (a program for the assessment of protein models with three-dimensional profiles). The resulting structure was further optimized through energy minimization before docking. The structures of the nine N-phenylpiperazine derivatives were constructed using DS 2.5 (Fig. 1). The docking study was carried out using the LIBDOCK program³² (site-features docking algorithm), implemented in the software platform of DS. The pose cluster radius was set to 0.5 with top hits of 10.

Fig. 1 Chemical structures of nine N-phenylpiperazine derivatives. Compound 1: 2-piperazin-4-amino-6,7-dimethoxy-quinazoline; compound 2: 1-(4-amino-6,7-dimethoxy-2-quinazolinyl)-4-(2-methyl-carbonyl)piperazine; compound 3: doxazosin; compound 4: prazosin; compound 5: terazosin; compound 6: 1-(4-amino-6,7-dimethoxy-2-quinazolinyl)-4-[(benzaldehyde 4-yl)carbonyl]piperazine; compound 7: 1-(2-furoyl carbamoyl)piperazine; compound 8: 1-acetyl-4-(2-furoyl carbamoyl)piperazine; compound 9: 1-acetyl-4-(2-tetrahydrofuryl carbamoyl)piperazine.

The docking results were characterized by the parameters of inhibition constant, hydrogen bond interaction energy, electrostatic energy, van der Waal’s forces and ligand efficiency. The conformation with the lowest energy was utilized in the docking process.

Frontal analysis

The binding mechanism of the N-phenylpiperazine derivatives to α_1A-AR was next analyzed via affinity chromatography using frontal analysis.³³ The capacity factors of each compound at 20 °C in the initial mobile phase (without addition of any compounds) were calculated using k′ = (t_R − t₀)/t₀, where k′ is the capacity factor, t_R is the retention time of the injection compound, and t₀ is the void time of the chromatographic system determined using NaNO₂ which is an un-retained solute on the column. The mobile phase was Tris-buffer (5.0 mM, pH 7.2) containing 1.0 mM NaCl and 0.5 mM EDTA. The flow rate was 0.2 mL min⁻¹ with detection wavelengths of 246 nm for doxazosin, prazosin and terazosin, and 254 nm for the other derivatives.

The association constants of the nine compounds for α_1A-AR were determined using frontal analysis. The mobile phases were solutions containing 0.25, 1.0, 2.0, 4.0, 6.0, 8.0, 12.0, 16.0 and 20.0 μM of each compound, prepared using the initial mobile phase. The breakthrough times of each concentration were triply determined to calculate the association constants of each compound for the receptor using eqn (1):

In this equation, K_A is the association equilibrium constant for the binding of analyte A to the immobilized ligand L, and m_Lapp is the apparent moles of analyte required to reach the mean point of a breakthrough curve at a given concentration of applied analyte [A]. According to eqn (1), the plot of 1/m_Lapp versus 1/[A] predicts a linear relationship when only one type of binding site is available on the column. The slope and intercept can be used to calculate K_A and the total moles of binding sites m_L for the analyte.³⁴

Results and discussion

Homology modeling of α_1A-AR

The construction of a protein model consists of four steps: sequence alignment between the target and the template, building an initial model, refining the model, and evaluating the quality of the model. In this work, these steps were followed well during the construction of the α_1A-AR homology model. Firstly, a template (3SN6, β₂-adrenergic receptor–Gs protein complex) was identified through homology searches in the PDB with a sequence identity of 53% to α_1A-AR (ESI†). Ten annotated structures of α_1A-AR were predicted, among which the desired model was Alpha.BL00020001 because of the good overlap with the template. This result was confirmed by the lowest value of PDF (Probability Density Functions) total energy (−29743.8), PDF physical energy (−5237.6) and DOPE (Discrete Optimized Protein Energy) score (−35692.4) compared with the other models. The Ramachandran plot of the α_1A-AR model, calculated using PROCHECK, showed that most of the amino acid residues of the receptor were distributed in the rational region (ESI†). Further verification of the model using PROFILE 3D showed that 78.4% of the residues had an averaged compatibility of an atomic model (3D) with its own amino acid sequence (1D).

Molecular docking analysis

It is reported that human α_1A-AR has 466 amino acids. Two phenylalanine residues (Phe²⁰⁸ and Phe³¹²) in transmembrane domain 7 (TM7), one phenylalanine residue (Phe¹⁹³) in TM5 and one leucine residue in TM6 (Leu²⁹⁰) have been identified as major sites for ligand binding by π-stacking and/or hydrophobic interactions.^35–37 Further experiments of alanine-substitution mutation have shown that two serine residues (Ser¹⁸⁸ and Ser¹⁹²) in TM5 play main role in the formation of hydrogen bonds between the receptor and ligands.³⁸ It is also reported that the protonated nitrogen of the bound ligand engages in ionic interactions with an aspartate residue (Asp¹⁰⁶) in TM3.^39,40 Other mutagenesis studies have indicated that Phe⁸⁶ in TM2 is capable of recognizing α_1A-AR selective antagonists as well as other dihydropyridine-type antagonists. The acidic amino residues Gln¹⁹⁶, Ile¹⁹⁷, and Asn¹⁹⁸ in the second extracellular loop discriminate the α_1A-AR selective antagonists (phentolamine and WB4101). Phe³⁰⁸ and Phe³¹² in TM7 are major aromatic contacts for most α₁-AR antagonists as well as imidazoline-type agonists.

The overview of the docking complexes of N-phenylpiperazine derivatives and α_1A-AR is presented in Fig. 2. Asp⁷⁸, Val⁷⁹, Cys⁸², Gln¹⁴⁹, Ile¹⁵⁰, Ser¹⁶⁰ and Ser¹⁶⁴ seemed to play important roles in the drug binding through hydrogen bond formation. The quinazoline and furan system mainly engaged in strong π-stacking and/or hydrophobic interactions with Phe⁵⁸ and Phe²⁵⁰. Other residues, including Arg⁶⁸, Trp⁷⁴, Tyr²⁵⁴, Trp²⁵¹, Ser⁵⁵, Cys¹⁴⁸, Trp⁶⁴, Thr¹⁴⁶, Asp⁷⁷, Ile¹⁴⁷ and Tyr⁶³, participated in ligand binding by van der Waal’s forces. The proposed functional group interactions, especially the binding involving Asp⁷⁸, Gln¹⁴⁹, Ser¹⁶⁰, Ser¹⁶⁴ and Phe¹⁶⁵, corresponded to the amino acid residues of Asp¹⁰⁶, Gln¹⁷⁷, Ser¹⁸⁸, Ser¹⁹² and Phe¹⁹³ in previous reports.³⁵ This means that the specific binding sites for the nine compounds to α_1A-AR were consistent with the results of mutagenesis studies.^36,37 In modelling the interactions of other compounds in this work, we concluded that antagonist binding was docked higher in the pocket than agonist binding, closer to the extracellular surface, and may be skewed toward TM7. This result was consistent with the previous modelling studies, where it was suggested that the α_1A-AR antagonists prazosin, tamsulosin and KMD-3213 docked with amino acid residues near the extracellular surface.

Fig. 2 An overview of the docking complexes of N-phenylpiperazine derivatives and α_1A-AR. The derivatives are depicted as sticks. Com 1–6 means compounds 1–6; Com 7–9 represents compounds 7–9.

Purification of α_1A-AR

As members of the G-protein coupled receptor superfamily, α-ARs have been sub-classified into several subtypes on the basis of their relative affinities for a variety of ligands. However, little progress has been made in the techniques for their solubilization and purification. In our previous work,²⁶ we synthesized a new affinity resin for the purification of α-AR by linking the highly specific antagonist, prazosin, on the surface of agarose gel. This proves that the resin has potential in the purification of native α₁-AR from lysates of cells or animal tissues. In this presentation, we have purified α_1A-AR using a commercial anti-FLAG M1 agarose column. Elution of the receptor was accomplished using antibody-mediated affinity chromatography in a calcium-dependent manner.

Compared with the method based on the prazosin-derived resin, this mild, calcium-dependent affinity chromatography procedure enabled rapid purification of α_1A-AR. Moreover, the entire purification was accomplished in a single step within several hours, starting from a crude cell homogenate or supernatant, without ever exposing the receptor to conditions other than physiological saline at pH 7.2 (with calcium or EDTA). It is notable that the limitation of the anti-Flag M1 monoclonal antibody is ascribed to its specificity only for the N-terminus of the FLAG fusion protein.

Determination of the association constants using frontal analysis

Affinity chromatography is one of the widely used techniques for exploring drug–protein interactions.^41,42 This set of experiments aimed to verify the molecular docking predicted mechanism via α_1A-AR affinity chromatography using frontal analysis. Representative chromatograms of prazosin from frontal analysis are depicted in Fig. 3 where increasing drug concentrations were used in the mobile phase. It was found that each profile of the chromatograms was in good agreement with the shape of a sigmoid curve. The same results were also found when the other compounds were applied in the mobile phases. These results indicated that frontal analysis was able to describe the adsorption and desorption behaviors of the nine derivatives on the stationary phase containing immobilized α_1A-AR.

Fig. 3 The representative chromatograms of prazosin from frontal analysis. The concentrations of the drug in the mobile phase were 0.25, 1.0, 2.0, 4.0, 6.0, 8.0, 12.0, 16.0 and 20.0 μM (bottom to top).

The plots obtained for 1/m_Lapp versus 1/[compound] are depicted in Fig. 4, which gave linear relationships with correlation coefficients ranging from 0.9950 to 0.9993 (Table 1). According to eqn (1), these results suggested that only a single type of binding site was responsible for the binding of the derivatives on the immobilized α_1A-AR. Table 2 summarizes the association constants of the compounds for α_1A-AR from frontal analysis. The affinity rank order of the nine compounds measured at 20 °C using frontal analysis was compound 3 > compound 4 > compound 6 > compound 5 > compound 2 > compound 1 > compound 7 > compound 8 > compound 9. This order was possible due to the variance in their structural properties.

Fig. 4 The plots of 1/m_Lapp versus 1/[compound] for the nine N-phenylpiperazine derivatives. ■, compound 1; □, compound 2; ●, compound 3; ○, compound 4; ▲, compound 5; ◆, compound 6; ◇, compound 7; ★, compound 8; ▼, compound 9. The experiments were performed at 20 °C.

Table 1 Regression equations of the curves from plotting 1/m_Lapp versus 1/[A] of the nine N-phenylpiperazine derivatives

Compound	Regression equation	Slope	Intercept	Correlation coefficient
1	1/mLapp = 0.42 × 10^−6/[Acom1] + 0.031	0.42	0.031	0.9977
2	1/mLapp = 0.46 × 10^−6/[Acom2] + 0.035	0.46	0.035	0.9981
3	1/mLapp = 0.25 × 10^−6/[Acom3] + 0.048	0.25	0.048	0.9993
4	1/mLapp = 0.39 × 10^−6/[Acom4] + 0.046	0.39	0.046	0.9966
5	1/mLapp = 0.47 × 10^−6/[Acom5] + 0.037	0.47	0.037	0.9983
6	1/mLapp = 0.34 × 10^−6/[Acom6] + 0.036	0.34	0.036	0.9984
7	1/mLapp = 0.64 × 10^−6/[Acom7] + 0.017	0.64	0.017	0.9981
8	1/mLapp = 0.66 × 10^−6/[Acom8] + 0.015	0.66	0.015	0.9950
9	1/mLapp = 0.58 × 10^−6/[Acom9] + 0.013	0.58	0.013	0.9969

---

Table 2 Calculation of the association constants of the N-phenylpiperazine derivatives binding to immobilized α_1A-AR

Compounds	Association constants (×10^5 M^−1)
	10 °C	20 °C	30 °C	37 °C	45 °C
1	0.80	0.74	0.66	0.62	0.58
2	0.82	0.76	0.68	0.64	0.59
3	2.04	1.92	1.76	1.68	1.59
4	1.29	1.18	1.05	0.98	0.92
5	0.84	0.79	0.75	0.71	0.68
6	1.12	1.05	0.98	0.95	0.92
7	0.32	0.27	0.23	0.21	0.19
8	0.26	0.23	0.21	0.19	0.18
9	0.25	0.22	0.19	0.17	0.16

---

A previous investigation using the pharmacophoric model⁴³ suggested that the three-dimensional structural properties of an ideal α₁-AR antagonist include the following: a positively ionizable group, corresponding to the more basic nitrogen atom of the aryl piperazine ring; an ortho or meta-substituted phenyl ring, both of which constitute the arylpiperazine system and satisfy three certain features of the pharmacophoric hypothesis; a polar group that provides a hydrogen bond acceptor feature, filling one of the portions of the pharmacophore that is required at the edge of the molecule opposite to the arylpiperazine moiety; and a hydrophobic moiety. The structures of compounds 7–9 only possessed a piperazine functional group and thus presented the approximate values of association constants which were far lower than the other six compounds. On the contrary, the structures of compounds 3–6 showed better agreement with these structural properties and as a result showed stronger affinity for α_1A-AR. Compound 9 presented the weakest affinity for the immobilized α_1A-AR due to it having little accord with the above-mentioned functional groups of ideal α_1A-AR ligands.

Focusing on the cases of prazosin and terazosin, the association constants determined using frontal analysis presented good agreement with the data from the literature under similar experimental conditions.^21,44 This comparison was rational since eqn (1) indicates that the determination of association constants by frontal analysis is independent of the number of immobilized proteins in column. This indication was possible because the values of the association constants were calculated from the ratio of the intercept to the slope of the equation. This was advantageous for comparison of the K_A values from the α_1A-AR columns with that from the column having different densities of the receptor. It was also valuable when precise measurement of association constants was required under the conditions where the number of immobilized receptors or the binding capacity of the column gradually decreased over the time.

Thermodynamic studies

To verify the binding mechanism of the nine compounds to α_1A-AR predicted by molecular docking, the thermodynamic behaviors during the interactions were investigated using affinity chromatography. In this investigation, the binding of the compounds to α_1A-AR were considered to be driven by weak intermolecular forces including hydrogen bonds, van der Waal’s forces, electrostatic forces and hydrophobic interactions. Assuming an identical value for enthalpy change (ΔH^θ) during the interaction, calculations of the enthalpy change (ΔH^θ), the entropy change (ΔS^θ) and the Gibbs’ free energy change (ΔG^θ) could be followed using eqn (2) and (3):

ΔG^θ = ΔH^θ − TΔS^θ (3)

where ΔH^θ and ΔS^θ describe the enthalpy and entropy changes accounting for the binding process of the compounds to the receptor, R is the ideal gas law constant and T is the absolute temperature. Using the semi-empirical law reported by Ross et al.,⁴⁵ the type of force that drives the binding interaction can be determined from the thermodynamic parameters in eqn (2) and (3).

When ΔH^θ > 0 and ΔS^θ > 0, the main force is considered to be hydrophobic interaction. Under the conditions that ΔH^θ < 0 and ΔS^θ > 0, electrostatic force is believed to be the main factor pushing the binding. When ΔH^θ < 0 and ΔS^θ < 0, hydrogen bond formation or van der Waal’s forces become the main forces during the interaction.

Table 2 lists the association constants of the nine compounds at 10, 20, 30, 37 and 45 °C, calculated using eqn (1). It was found that the association constants decreased with growing temperature, while the number of binding sites showed a positive correlation with the increasing temperature. This result was rational because the receptor has at least two conformations at the initial time when the column is prepared. Along with the growth in the temperature, a number of receptors at the rest state changed their conformation to the active state, and then served as the binding site for the compounds to α_1A-AR. It should be noticed that the capacity factors of the compounds on the immobilized α_1A-AR decreased with increasing temperature. These results indicated that the retention behavior of the compounds is ascribed to the synergistic contribution of the binding site and the association constant. The affinity was the main factor that determined the retention behavior.

The results in Table 2 were further utilized to uncover the relationship between the association constants and the temperature. As predicated in eqn (2), good linear relationships were obtained between ln K_A and 1000/T for all the compounds in the given temperature range (Fig. 5).

Fig. 5 The plots of ln K_A versus 1000/T for the nine N-phenylpiperazine derivatives. ■, compound 1; □, compound 2; ●, compound 3; ○, compound 4; ▲, compound 5; ◆, compound 6; ◇, compound 7; ★, compound 8; ▼, compound 9.

As summarized in Table 3, all the compounds gave a principle of ΔH^θ < 0, ΔS^θ > 0 and ΔG^θ < 0, which indicated an endothermic process with an increase in entropy for the interaction. According to the semi-empirical law, the driving force for this process was electrostatic interaction. This result was reasonable because all the compounds were used in the form of amine salts. For instance, prazosin is widely used as prazosin hydrochloride.

Table 3 Thermodynamic parameters of the N-phenylpiperazine derivatives binding to immobilized α_1A-AR, calculated at 20 °C

Compounds	Regression equation	Correlation coefficients	ΔH^θ kJ mol^−1	ΔS^θ J mol^−1 K^−1	ΔG^θ kJ mol^−1
1	ln KA = 848.7/T + 8.30	0.9950	−7.06 ± 0.24	69.01 ± 2.16	−27.29
2	ln KA = 856.4/T + 8.30	0.9934	−7.12 ± 0.18	69.01 ± 3.02	−27.35
3	ln KA = 653.7/T + 9.92	0.9955	−5.43 ± 0.12	82.47 ± 4.18	−29.61
4	ln KA = 893.3/T + 8.62	0.9966	−7.43 ± 0.31	71.26 ± 1.47	−28.32
5	ln KA = 545.5/T + 9.42	0.9950	−4.54 ± 0.26	78.32 ± 2.69	−27.50
6	ln KA = 516.1/T + 9.80	0.9933	−4.29 ± 0.15	81.47 ± 1.94	−28.17
7	ln KA = 1347/T + 5.61	0.9988	−11.2 ± 0.11	46.64 ± 2.07	−24.87
8	ln KA = 960.1/T + 6.77	0.9943	−7.98 ± 0.10	56.29 ± 1.86	−24.48
9	ln KA = 1194/T + 5.91	0.9937	−9.93 ± 0.22	49.14 ± 2.41	−24.34

---

Correlation between the fit scores and association constant

In molecular docking, the fit score should give a positive correlation with the affinity of a drug for a receptor. In this vein, the K_A values determined using an experimental method are expected to correlate with the docking scores. In this work, the fit scores of the nine compounds were determined to be 93.4 for compound 1, 95.6 for compound 2, 123.2 for compound 3, 117.3 for compound 4, 98.9 for compound 5, 103.8 for compound 6, 87.9 for compound 7, 83.6 for compound 8 and 80.2 for compound 9. The rank order of the scores for the nine compounds binding to α_1A-AR was: compound 3 > compound 4 > compound 6 > compound 5 > compound 2 > compound 1 > compound 7 > compound 8 > compound 9. This was in good agreement with the pattern of the association constants from frontal analysis.

To further reveal the relationship between the fit scores from molecular docking and the K_A values from frontal analysis, we plotted the curve of the scores versus the association constants using linear analysis. The result presented a linear relationship between the fit scores and the association constants (Fig. 6). The regression equation was y = 23.1x + 77.5 with a correlation coefficient of 0.9860. This agreement indicated that the proposed HPAC method will probably become an alternative for exploring drug–receptor binding mechanisms.

Fig. 6 The plot of the fit scores from molecular docking versus the association constant from frontal analysis.

Conclusions

The binding mechanism of nine N-phenylpiperazine derivatives to α_1A-AR was investigated using molecular docking and frontal analysis. It was found that the binding site of the nine derivatives to the receptor was located at the amino acid residues of Gln¹⁴⁹, Phe²⁵⁰ Asp⁷⁷ and Ser⁵⁵. Both molecular docking and thermodynamic investigation by frontal analysis showed that the interaction between N-phenylpiperazine derivatives and α_1A-AR was driven by electrostatic forces. The molecular docking technique is capable of predicating the mechanism of drug–protein interaction. The immobilized receptor stationary phase has the capacity to elucidate the drug–receptor binding mechanism and will probably become a powerful methodology for the design and screening of drug candidates specifically binding to a receptor.

Acknowledgements

We are thankful to Professor Jiang Ru in the Fourth Military University for providing the DS software platform and for the financial support from the grant of the National Natural Science Foundation of China (no. 21475103), the Natural Science Foundation of Shaanxi Province (no. 2015JM2072), the program for Innovative Research Team of Shaanxi Province (no. 2013KCT-24) and the Ministry of Science and Technology of the People’s Republic of China (no. 2013YQ170525; subproject: 2013YQ17052509).

Notes and references

1. S. Schilit and K. E. Benzeroual, Clin. Ther., 2009, 31, 2489–2502.
2. B. C. Jensen, T. D. O’Connell and P. C. Simpson, J. Mol. Cell. Cardiol., 2011, 51, 518–528.
3. P. Kolb, D. M. Rosenbaum, J. J. Irwin, J. J. Fung, B. K. Kobilk and B. K. Shoichet, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 6843–6848.
4. I. Nalepa, G. Kreiner, A. Bielawski, K. Rafa-Zablocka and A. Roman, Pharmacol. Rep., 2013, 65, 1489–1497.
5. J. Hwa, R. M. Graham and D. M. Perez, J. Biol. Chem., 1995, 270, 23189–23195.
6. S. Cotecchia, J. Recept. Signal Transduction, 2010, 30, 410–419.
7. J. G. Baker, Br. J. Pharmacol., 2010, 160, 1048–1061.
8. C. Hosoda, T. Koshimizu, A. Tanoue, Y. Nasa, R. Oikawa, T. Tomabechi, S. Fukuda, H. Shinoura, S. Oshikawa, S. Takeo, T. Kitamura, S. Cotecchia and G. Tsujimoto, Mol. Pharmacol., 2005, 67, 912–922.
9. R. M. Graham, D. M. Perez, J. Hwa and M. T. Piascik, Circ. Res., 1996, 78, 737–749.
10. I. M. Kapetanovic, Chem.-Biol. Interact., 2008, 171, 165–176.
11. R. Macarron, M. N. Banks, D. Bojanic, D. J. Burns, D. A. Cirovic, T. Garyantes, D. V. S. Green, R. P. Hertzberg, W. P. Janzen, J. W. Paslay, U. Schopfer and G. S. Sittampalam, Nat. Rev. Drug Discovery, 2011, 10, 188–195.
12. C. W. Murray and D. C. Rees, Nat. Chem., 2009, 1, 187–192.
13. S. G. Patching, Biochim. Biophys. Acta, 2014, 1838, 43–55.
14. S. J. Sun, C. J. Long, C. Y. Tao, S. Meng and B. Y. Deng, Anal. Chim. Acta, 2014, 851, 37–42.
15. I. Vayá, R. Pérez-Ruiz, V. Lhiaubet-Vallet, M. C. Jiménez and M. A. Miranda, Chem. Phys. Lett., 2010, 486, 147–153.
16. R. Matsuda, S. H. Kye, J. Anguizola and D. S. Hage, Rev. Anal. Chem., 2014, 33, 79–94.
17. K. Jozwiak, S. Ravichandran, J. R. Collins and I. W. Wainer, J. Med. Chem., 2004, 47, 4008–4021.
18. J. Z. Chen and D. S. Hage, Nat. Biotechnol., 2004, 22, 1445–1448.
19. R. Moaddel, A. Rosenberg, K. Spelman, J. Frazier, C. Frazier, S. Nocerino, A. Brizzi, C. Mugnaini and I. W. Wainer, Anal. Biochem., 2011, 412, 85–91.
20. X. Y. Wang, X. Meng, W. J. Pei, Z. J. Li, Y. Y. Zhang, J. B. Zheng and X. H. Zheng, Anal. Methods, 2012, 4, 3420–3424.
21. X. F. Zhao, H. Y. Lu, J. J. Huang, J. B. Zheng, X. H. Zheng and Y. Y. Zhang, Chromatographia, 2012, 75, 411–415.
22. H. Marona, M. Kubacka, B. Filipek, A. Siwek, M. Dybała, E. Szneler, T. Pociecha, A. Gunia and A. M. Waszkielewicz, Pharmazie, 2011, 66, 733–739.
23. H. F. Wang, Z. L. Zuo, Q. L. Yang, X. J. Shen, X. F. Zhao and X. H. Zheng, Chin. J. Synth. Chem., 2010, 18, 745–746.
24. Y. Wang, B. X. Yuan, X. L. Deng, L. C. He, Y. Y. Zhang and Q. D. Han, Anal. Biochem., 2005, 339, 198–205.
25. A. Sousa, D. Bicho, C. T. Tomaz, F. Sousa and J. A. Queiroz, J. Chromatogr. A, 2011, 1218, 1701–1706.
26. X. F. Zhao, Q. Li, L. J. Bian, X. H. Zheng, J. B. Zheng, Y. Y. Zhang and Z. J. Li, J. Pharm. Biomed. Anal., 2012, 70, 549–552.
27. S. G. F. Rasmussen, B. T. deVree, Y. Z. Zou, A. C. Kruse, K. Y. Chung, T. S. Kobilka, F. S. Thian, P. S. Chae, E. Pardon, D. Calinski, J. M. Mathiesen, S. T. A. Shah, J. A. Lyons, M. Caffrey, S. H. Gellman, J. Steyaert, G. Skiniotis, W. I. Weis, R. K. Sunahara and B. K. Kobilka, Nature, 2011, 477, 549–555.
28. A. O’Driscoll, V. Belogrudov, J. Carroll, K. Kropp, P. Walsh, P. Ghazal and R. D. Sleator, J. Biomed. Inf., 2015, 54, 58–64.
29. R. Nair and B. Rost, Protein Sci., 2002, 11, 2836–2847.
30. R. A. Laskowski, M. W. MacArthur, D. S. Moss and J. M. Thornton, J. Appl. Crystallogr., 1993, 26, 283–291.
31. M. J. Thompson, S. A. Sievers, J. Karanicolas, M. I. Ivanova, D. Baker and D. Eisenberg, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 4074–4078.
32. S. N. Rao, M. S. Head, A. Kulkarni and J. M. LaLonde, J. Chem. Inf. Model., 2007, 47, 2159–2171.
33. S. Claesson, Discuss. Faraday Soc., 1949, 7, 34–38.
34. D. S. Hage, A. Jackson, M. R. Sobansky, J. E. Schiel, M. J. Yoo and K. S. Joseph, J. Sep. Sci., 2009, 32, 835–853.
35. L. P. Du and M. Y. Li, Curr. Comput.-Aided Drug Des., 2010, 6, 165–178.
36. D. J. Waugh, R. J. Gaivin, M. J. Zuscik, P. Gonzalez-Cabrera, S. A. Ross, J. Yun and D. M. Perez, J. Biol. Chem., 2001, 276, 25366–25371.
37. J. M. Wetzel, J. A. Salon, J. A. Tamm, C. Forray, D. Craig, H. Nakanishi, W. Cui, P. J. Vaysse, G. Chiu, R. L. Weinshank, P. R. Hartig, T. A. Branchek and C. Gluchowski, Recept. Channels, 1996, 4, 165–177.
38. D. F. McCune, R. J. Gaivin, B. R. Rorabaugh and D. M. Perez, Recept. Channels, 2004, 10, 109–116.
39. J. Hwa and D. M. Perez, J. Biol. Chem., 1996, 271, 6322–6327.
40. D. M. Perez, Biochem. Pharmacol., 2007, 73, 1051–1062.
41. X. F. Zhao, Q. Li, J. J. Chen, C. N. Xiao, L. J. Bian, J. B. Zheng, X. H. Zheng, Z. J. Li and Y. Y. Zhang, J. Chromatogr. A, 2014, 1339, 137–144.
42. X. F. Zhao, Q. Li, C. N. Xiao, Y. J. Zhang, L. J. Bian, J. B. Zheng, X. H. Zheng, Z. J. Li, Y. Y. Zhang and T. P. Fan, Anal. Bioanal. Chem., 2014, 406, 2975–2985.
43. L. Betti, M. Botta, F. Corelli, M. Floridi, G. Giannaccini, L. Maccari, F. Manetti, G. Strappaghetti, A. Tafi and S. Corsano, J. Med. Chem., 2002, 45, 3603–3611.
44. X. K. Gao, Y. Z. Li, Y. Qin, E. H. Chen, Q. Li, X. F. Zhao, L. J. Bian, J. B. Zheng, Z. J. Li, Y. Y. Zhang and X. H. Zheng, RSC Adv., 2015, 5, 24449–24454.
45. P. D. Ross and S. Subramanian, Biochemistry, 1981, 20, 3096–3102.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra10812h

---