Abstract
Species isoforms of histamine H2-, H3-, and H4-receptors differ in their pharmacological properties. The study aim was to dissect differences between the human H1R (hH1R) and guinea pig H1R (ghH1R). We coexpressed hH1R and gpH1R with regulators of G-protein signaling in Sf9 insect cells and analyzed the GTPase activity of Gq-proteins. Small H1R agonists showed similar effects at hH1R and gpH1R, whereas bulkier 2-phenylhistamines and histaprodifens were up to ∼10-fold more potent at gpH1R than at hH1R. Most 2-phenylhistamines and histaprodifens were more efficacious at gpH1R than at hH1R. Several first-generation H1R antagonists were ∼2-fold, and arpromidine-type H1R antagonists up to ∼10-fold more potent at gpH1R than at hH1R. [3H]Mepyramine competition binding studies confirmed the potency differences of the GTPase studies. Phe-153→Leu-153 or Ile-433→Val-433 exchange in hH1R (hH1R→gpH1R) resulted in poor receptor expression, low [3H]mepyramine affinity, and functional inactivity. The Phe-153→Leu-153/Ile-433→Val-433 double mutant expressed excellently but only partially changed the pharmacological properties of hH1R. Small H1R agonists and 2-phenylhistamines interacted differentially with human and guinea pig H2R in terms of potency and efficacy, respectively. Our data show the following: 1) there are differences in agonist- and antagonist-pharmacology of hH1R and gpH1R encompassing diverse classes of bulky ligands. These differences may be explained by higher conformational flexibility of gpH1R relative to hH1R; 2) Phe-153 and Ile-433 are critical for proper folding and expression of hH1R; and 3) H2R species isoforms distinguish between H1R agonists.
Histamine serves as a neurotransmitter and autacoid and acts through specific HxRs designated as H1R, H2R, H3R, and H4R, respectively (Hill et al., 1997; Hough, 2001). The H1R couples to Gq-proteins. Numerous H1R agonists and antagonists are known. H1R agonists are divided into three classes (Fig. 1): 1) small agonists (2–4) derived from histamine (1), 2) histamine derivatives with bulkier aromatic substituents at position 2 of the imidazole ring (5–18), and 3) histaprodifens, e.g., compounds 19 to 23 (Leschke et al., 1995; Zingel et al., 1995; Elz et al., 2000). H1R agonists are important experimental tools to analyze H1R function in cellular and organ systems (Zingel et al., 1995; Hill et al., 1997). H1R antagonists are commonly divided into sedating (first-generation, 24–32) and nonsedating (second-generation, 41–45) antagonists (Fig. 2). Today, especially the second-generation H1R antagonists are of great importance for the treatment of allergic diseases (Hill et al., 1997). Guanidines 33, 34, and 36 to 39 derived from arpromidine (35) are dual H2R agonists/H1R antagonists (Buschauer, 1989).
The availability of HxR cDNAs allowed for the comparison of the pharmacological properties of HxR species isoforms in recombinant systems under identical experimental conditions. Such expression studies uncovered species differences in the pharmacological properties of hH2R and gpH2R (Kelley et al., 2001), rat and human H3R (Ligneau et al., 2000; Lovenberg et al., 2000), and H4R from mouse, rat, guinea pig, and humans (Liu et al., 2001). Species differences in the pharmacological properties of HxRs provided opportunities to analyze the molecular basis of ligand/GPCR interactions (Ligneau et al., 2000; Kelley et al., 2001). From the standpoint of drug design, the pharmacological properties of hHxRs are important because in the HxR field essentially all the structures generated so far were derived from animal models, mostly from rat and guinea pig (Zingel et al., 1995; Hill et al., 1997).
The species differences in pharmacological properties of H2R, H3R, and H4R raise the question whether this is a general characteristic of HxRs. In fact, the Kd values of [3H]mepyramine for H1Rs from various species differ by ∼2 to 6-fold (Chang et al., 1979). Moreover, histaprodifens exhibit different potencies and efficacies in the guinea pig ileum and rat aorta (Elz et al., 2000). Furthermore, 2-(3-chlorophenyl)histamine (12) is a potent H1R agonist in the guinea pig ileum but failed to exhibit agonistic activity in H1R-expressing dibutyryl cAMP-differentiated human HL-60 leukemia cells (Seifert et al., 1994). A snake plot of hH1R depicts the relative positions and topology of amino acid residues in the TM domains, putative agonist and antagonist binding sites, and differences with respect to the gpH1R (Fig. 3). Mutagenesis data (Leurs et al., 1994, 1995; Ohta et al., 1994; Nonaka et al., 1998) and modeling approaches (Elz et al., 2000) indicated that histamine and histaprodifens interact with amino acid residues in TMs III, IV, V, and VII. Considering the alignment of H1Rs with bovine rhodopsin (Palczewski et al., 2000) and results of the substituted-cysteine accessibility method with the dopamine D2-receptor (Ballesteros et al., 2001), there are no amino acid differences in the ligand binding pocket of gpH1R and hH1R. The two lipid-directed residues, Phe-153 in TM IV of hH1R versus Leu in gpH1R and Ile-433 in TM VI of hH1R versus Val in gpH1R, represent the only differences near the binding site. Although these amino acid exchanges are conservative, the amino acids in hH1R are bulkier than those in gpH1R, and such differences could have an impact on the ligand-binding pocket.
The aim of the present study was to compare recombinant hH1R and gpH1R expressed in Sf9 insect cells under identical experimental conditions. We also examined the roles of Phe-153 and Ile-433 in hH1R function. As read-out, we focused on the determination of the GTPase activity of insect cell Gq-proteins in the presence of the RGS proteins RGS4 and GAIP. This coexpression system provides a sensitive model for studying H1R at the G-protein level (Houston et al., 2002). The GTPase assay is a steady-state method and eliminates the impact of effector availability/compartmentation and pharmacokinetic barriers on the properties of agonists (Buschauer, 1989; Ostrom et al., 2000). Moreover, we conducted [3H]mepyramine binding studies and analyzed the effects of H1R agonists on recombinant H2R-Gsα fusion proteins, recently verified as sensitive systems for the analysis of H2Rs (Kelley et al., 2001).
Materials and Methods
Materials. Construction of the cDNAs for hH1R-F153L, hH1R-I433V, and hH1R-F153L/I433V was performed by overlap-extension polymerase chain reaction following recently described procedures (Houston et al., 2002). Dimethindene enantiomers were a kind gift of Dr. G. Lambrecht (Department of Pharmacology, University of Frankfurt/Main, Germany). Ketotifen was a gift from Novartis (Basel, Switzerland), azelastine a gift from Asta Medica (Frankfurt/Main, Germany), fexofenadine a gift from Janssen-Cilag (Neuss, Germany), and terfenadine a gift from Aventis (Frankfurt/Main, Germany). Guanidines 33 to 38 were synthesized as described (Buschauer, 1989). Guanidine 39 was prepared by analogy to the procedures described for guanidines 33 to 38. 2-Methylhistamine (2) and 2-(2-thiazolyl)ethanamine (3) were synthesized using standard procedures. Compounds 5 to 18 were prepared according to published procedures (Zingel et al., 1990; Leschke et al., 1995). Compounds 22, 23, and 40 were available by synthetic pathways reported for the synthesis of 19 to 21 (Elz et al., 2000). Structures of synthesized compounds were confirmed by elemental analysis (C, H, and N), 1H NMR spectroscopy, and mass spectrometry. The purity of the compounds was >98% as determined by high-performance liquid chromatography or capillary electrophoresis. Tunicamycin, histamine, betahistine, promazine, chlorpromazine, mianserin, cyproheptadine, diphenhydramine, mepyramine, triprolidine, and (+)-chlorpheniramine were from Sigma-Aldrich (St. Louis, MO). Sources of other materials are described elsewhere (Kelley et al., 2001; Houston et al., 2002).
Cell Culture and Membrane Preparation. Recombinant baculoviruses encoding hH1R-F153L, hH1R-I433V, and hH1R-F153L/I433V were generated in Sf9 cells using the BaculoGOLD transfection kit (BD Pharmingen, San Diego, CA), according to the manufacturer's instructions. Infection and culture of Sf9 cells and membrane preparation were performed as described (Kelley et al., 2001; Houston et al., 2002). In some cultures, we added tunicamycin (10 μg/ml) to cultures to inhibit N-glycosylation of H1Rs (Seifert and Wenzel-Seifert, 2001).
[3H]Mepyramine Binding Assay. Membranes expressing various H1R constructs plus RGS proteins were thawed and sedimented by a 15-min centrifugation at 4°C and 15,000g. Membranes were resuspended in binding buffer (12.5 mM MgCl2, 1 mM EDTA, and 75 mM Tris/HCl, pH 7.4). Tubes (total volume 500 μl) contained 20–25 μg of membrane protein. Incubations were conducted for 90 min at 25°C and shaking at 250 rpm. For H1R saturation binding experiments, tubes contained 0.2 to 20 nM [3H]mepyramine (hH1R, gpH1R, and hH1R-F153L/I433V) or 2 to 100 nM [3H]mepyramine (hH1R-F153L and hH1R-I433V). Nonspecific binding was routinely determined in the presence of 10 μM mepyramine (30). Nonspecific binding in the presence of saturating concentrations of compounds 1, 3, 12, 14, 15, 19, 20, 31, 35, and 36 was virtually identical to nonspecific binding in the presence of compound 30 (data not shown). Competition binding experiments were carried out in the presence of 2 nM [3H]mepyramine and unlabeled ligands at various concentrations. Bound [3H]mepyramine was separated from free [3H]mepyramine by filtration through GF/C filters, followed by three washes with 2 ml of binding buffer (4°C). Filter-bound radioactivity was determined by liquid scintillation counting.
Steady-State GTPase Activity Assay. Membranes expressing various H1R constructs plus RGS proteins or H2R-Gsα fusion proteins were thawed, sedimented, and resuspended in 10 mM Tris/HCl, pH 7.4. Assay tubes contained Sf9 membranes (10 μg of protein/tube), 1.0 mM MgCl2, 0.1 mM EDTA, 0.1 mM ATP, 100 nM GTP, 1 mM adenylyl imidodiphosphate, 5 mM creatine phosphate, 40 μg of creatine kinase, and 0.2% (w/v) bovine serum albumin in 50 mM Tris/HCl, pH 7.4, and HxR ligands at various concentrations. Reaction mixtures (80 μl) were incubated for 3 min at 25°C before the addition of 20 μl of [γ-32P]GTP (0.2–0.5 μCi/tube). Reactions were conducted for 20 min at 25°C. Reactions were terminated by the addition of 900 μl of slurry consisting of 5% (w/v) activated charcoal and 50 mM NaH2PO4, pH 2.0. Charcoal-quenched reaction mixtures were centrifuged for 15 min at room temperature at 15,000g. Seven hundred microliters of the supernatant fluid of reaction mixtures were removed, and 32Pi was determined by liquid scintillation counting.
SDS-PAGE and Immunoblot Analysis. Membrane proteins were separated on SDS polyacrylamide gels containing 10% (w/v) acrylamide. Proteins were then transferred onto Immobilon-P transfer membranes (Millipore, Bedford, MA). Membranes were reacted with M1 antibody (1:1000). Immunoreactive bands were visualized by sheep anti-mouse IgG (1:1000) coupled to peroxidase, using o-dianisidine and H2O2 as substrates. Expression of RGS proteins was verified by immunoblot analysis with specific anti-RGS4 IgG and anti-GAIP IgG, as described (Houston et al., 2002).
Miscellaneous. Protein concentrations were determined using the Bio-Rad DC protein assay kit (Bio-Rad, Hercules, CA). All analyses of experimental data were performed with the Prism 3.02 software (GraphPad-Prism, San Diego, CA). Ki and KB values were calculated according to Cheng and Prusoff (1973). Statistical comparisons were performed with Student's t test.
Results
Immunological Detection of H1R Constructs. The H1R constructs analyzed in this study were all N-terminally tagged with the FLAG epitope, allowing immunological detection with the M1 monoclonal antibody (Houston et al., 2002). The predicted molecular mass of nonglycosylated hH1R and gpH1R is ∼56 kDa (Fukui et al., 1994; Traiffort et al., 1994). The FLAG epitope-tagged hH1R expressed in Sf9 membranes migrated as diffuse ∼75-kDa doublet in SDS-PAGE (Figs. 4, A and B). Treatment of Sf9 cells with the inhibitor of N-glycosylation, tunicamycin (Seifert and Wenzel-Seifert, 2001), shifted the majority of the protein toward 70 kDa and rendered the lower band crisper. Migration of FLAG epitope-tagged gpH1R in SDS-PAGE differed considerably from the migration of hH1R. In membranes expressing gpH1R, faint and diffuse bands in the ∼36- and ∼50-kDa regions were detected, and tunicamycin treatment had little effect on migration of gpH1R in SDS-PAGE (Fig. 4A). Additionally, we detected intense and crisp bands of ∼16 and ∼30 kDa. Both hH1R-F153L and hH1R-I433V showed a broad ladder of diffuse bands ranging from ∼30 to 80 kDa, and there was a more intense doublet at 28 to 29 kDa (Fig. 4B). The hH1R-F153L/I433V double mutant showed the predicted migration in SDS-PAGE, i.e., this mutant migrated as a ∼56 kDa band.
Analysis of H1R Constructs in [3H]Mepyramine Binding Assays. The Kd of [3H]mepyramine for hH1R expressed in Sf9 membranes was 1.8-fold higher than the Kd for gpH1R (Table 1). The Bmax values of hH1R and gpH1R expression in Sf9 membranes were similar to the expression levels reported for the β2-adrenoceptor (Seifert et al., 1998). Compared with hH1R, the [3H]mepyramine-affinities of hH1R-F153L and hH1R-I433V were reduced by ∼8- to 12-fold, and the Bmax values were reduced by ∼5- to 6-fold. The double mutation restored [3H]mepyramine-affinity of hH1R and efficient expression.
Potencies and Efficacies of H1R and H2R Agonists at H1R Constructs in the GTPase Assay. We studied three classes of H1R agonists in the GTPase assay (Fig. 1). As a control we also studied the H2R agonists amthamine (46) and dimaprit (47) (Hill et al., 1997). Table 2 and Fig. 5 summarize the data for hH1R and gpH1R coexpressed with RGS4 and GAIP since no significant differences were observed between the two RGS proteins (data not shown). Only histamine and the small histamine derivatives 2 and 3 were full hH1R agonists, whereas all other modifications resulted in reductions of efficacy. Additionally, compounds 2 and 3 were less potent hH1R agonists than histamine. We identified only two agonists that were more potent at hH1R than histamine, i.e., the histaprodifens 19 and 20. The moderate increase in potency (1.8–2.7-fold) was accompanied by a significant decrease in efficacy, however. The introduction of a phenyl group (6) or particularly a benzyl group (5) at the position 2 of the imidazole ring substantially reduced agonist potency. Introduction of a halogen in the meta-position of the phenyl ring partially restored agonist potency in the order F < Cl < Br ∼ I (compare 6, 9, 12, 14, and 15). Other hydrogen-donating meta-substituents (Oe and CF3) were also favorable (16 and 17), whereas a methyl group (7) and halogen substitutions in the ortho- or para-position of the phenyl ring (8 and 13) further reduced agonist potency. At hH1R, histaprodifens 21 to 23 were less potent than histamine. The H2R agonists 46 and 47 were essentially devoid of agonistic activity at the hH1R (Table 2).
We did not observe significant differences in potency and efficacy of the small H1R agonists 1 to 4 between hH1R and gpH1R (Table 2). This similarity between the H1R isoforms is reflected by a linear correlation of the pEC50 values of the small agonists at hH1R and gpH1R that is close to the theoretical correlation describing identity of H1R species isoforms (Fig. 5A). When the effects of 2-phenylhistamines and histaprodifens were analyzed, however, significant differences between hH1R and gpH1R emerged. All compounds of these two classes were significantly more potent (3.2- to 9.9-fold) at gpH1R than at hH1R. The different interaction of 2-phenylhistamines and histaprodifens with hH1R and gpH1R is reflected by a linear correlation of the potencies of each series that is shifted toward the left relative to the theoretical correlation describing pharmacological identity of the GPCR species isoforms (Figs. 5, B and C). These linear correlations also show that the overall structure/activity relationships of those compounds are similar at both H1R species isoforms. In addition to the higher potency, most 2-phenylhistamines (6, 8–12, 14–17) and 3 of 5 histaprodifens (20, 21, and 23) were significantly more efficacious at gpH1R than at hH1R. Finally, the small H2R agonist dimaprit (47) showed only minimal agonistic effects at gpH1R, but another small agonist, amthamine (46), was a weak partial gpH1R agonist with significantly higher efficacy at gpH1R than at hH1R.
We failed to detect GTPase stimulation by histamine and compounds 3 and 12 in Sf9 membranes expressing hH1R-F153L and hH1R-I433V plus RGS proteins (data not shown). In contrast, histamine and compound 3 stimulated GTP hydrolysis in membranes expressing hH1R-F153L/I433V as potently and efficiently as in membranes expressing hH1R or gpH1R. 2-Substituted histamines and histaprodifens tended to be more potent and efficacious at hH1R-F153L/I433V than at hH1R, but only the potency and efficacy of compound 12 were significantly increased.
Constitutive Activity of H1Rs. hH1R is constitutively active, and many first- and second-generation H1R antagonists possess inverse agonistic activity (Bakker et al., 2001; Weiner et al., 2001). The extent of constitutive activity of hH1R is dependent on the specific expression system, however. All first-generation H1R antagonists (24–32), second-generation H1R antagonists (41–45), and guanidines (33–39) examined exhibited only small inverse agonistic activity at hH1R expressed in Sf9 membranes, i.e., the inhibitory effects of compounds amounted to ∼5 to 15% of the stimulatory effect of histamine (data not shown). There were no significant differences in the inverse agonist effects of H1R antagonists at hH1R and gpH1R. These data indicate that the constitutive activity of the two GPCR isoforms is similar.
Potencies of H1R Antagonists at H1R Constructs in the GTPase Assay. In agreement with the [3H]mepyramine binding studies (Table 1), mepyramine (30) was about 2-fold less potent at inhibiting histamine-stimulated GTP hydrolysis in membranes expressing hH1R than in membranes expressing gpH1R (Table 3). A similar difference in potency was observed for two other first-generation H1R antagonists, triprolidine (31) and (+)-chlorpheniramine (32), whereas the other first-generation antagonists studied [24–28, dimethindene enantiomers (R)-(-)-29 and (S)-(+)-29] did not exhibit significantly different potencies at hH1R and gpH1R. (R)-(-)-Dimethindene was ∼30- to 40-fold more potent than (S)-(+)-dimethindene. The stereoselectivity of recombinant H1Rs for dimethindene enantiomers is in accordance with data for the H1R expressed in the guinea pig ileum (Pfaff et al., 1995). Among the second-generation H1R antagonists 41 to 45, no significant differences in potency between hH1R and gpH1R emerged.
Arpromidine (35) and arpromidine-derived guanidines (33, 34, and 36–38) are not only very potent H2R agonists but also moderately potent H1R antagonists (Buschauer, 1989). The H1R-antagonistic properties of guanidines are explained by the structural similarity of compounds 33 to 38 and 30 to 32 (Fig. 2). Guanidines 33 to 38 inhibited histamine-stimulated GTP hydrolysis in Sf9 membranes expressing gpH1R, with KB values of ∼50 to 150 nM (Table 3). Guanidines 33 to 38 were all significantly more potent antagonists at gpH1R than at hH1R and showed greater gpH1R/hH1R selectivity than compounds 30 to 32. The difference in potency was most pronounced (∼9-fold) for compound 36 that is distinguished from the other guanidines by a para-Cl in the phenyl moiety (Fig. 2). In contrast, guanidine 39 that possesses a trichlorinated phenyl ring and a thiazole instead of a pyridyl ring (Fig. 2) did not discriminate between hH1R and gpH1R. Modifications of the substituents in guanidines 33–39 had a considerably larger impact on antagonist potency at gpH1R (∼7-fold) than at hH1R (∼2-fold).
In the 2-phenylhistamine derivative 40, the free amino group of histamine was integrated into a piperidine ring (Fig. 2). This modification is predicted to interfere with the binding of the basic nitrogen to Asp-107 (hH1R) (Ohta et al., 1994). In fact, compound 40 exhibited 6.5- to 8-fold reduced apparent affinity compared with its parent compound (17) (Fig. 1) at hH1R and gpH1R (Tables 2 and 3). Moreover, introduction of the piperidine ring into 17 conferred antagonistic properties to compound 40 (Table 3). This was also confirmed in the guinea pig ileum assay (KB of compound 40, 400 nM). Compound 40 was a severalfold more potent antagonist at gpH1R than at hH1R.
In agreement with the binding data (Table 1), mepyramine (30) was similarly potent at inhibiting histamine-stimulated GTP hydrolysis in Sf9 membranes expressing hH1R and hH1R-F153L/I433V (Table 3). The double mutation exhibited inconsistent effects on the potencies of guanidines 33 and 35 to 38 as well as of the 2-phenylhistamine derivative 40. Specifically, the F153L/I433V mutation increased the potency of 36 1.5-fold, had no effect on the potency of 35, and 37 and decreased the potency of compounds 33, 38, and 40 by up to 2-fold.
Affinities of H1R Agonists and Antagonists at H1R Constructs in the [3H]Mepyramine Binding Assay. Histamine and 2-(3-chlorophenyl)histamine (12) inhibited [3H]mepyramine binding in Sf9 membranes expressing hH1R or gpH1R plus RGS proteins according to a monophasic function that was not shifted to the right by guanosine 5′-O-(3-thiotriphosphate) (10 μM) (data not shown). Thus, we could not detect high-affinity agonist binding. These data were expected since there is a paucity of endogenous G-proteins relative to the expressed mammalian GPCRs in Sf9 membranes (Seifert et al., 1998; Houston et al., 2002). Accordingly, the agonist-affinities determined in the [3H]mepyramine competition binding studies reflect the agonist affinities of H1Rs in the G-protein-uncoupled state. In fact, the Ki values of agonists 1, 3, 12, 14, 15, 19, and 20 at hH1R and gpH1R were all higher than the corresponding EC50 values in the GTPase assay (Tables 2 and 4). The Ki value of histamine at hH1R was 2.3-fold lower than the Ki value of histamine at gpH1R. Since the amino acids in the histamine-binding H1R domains are identical in both isoforms (Fig. 3), this difference could point to a better fit of histamine into the Gq-uncoupled hH1R compared with Gq-uncoupled gpH1R.
To account for the difference in histamine affinity of H1R species isoforms, we focused on the comparison of the relative affinities of synthetic agonists at hH1R and gpH1R. The relative affinity of the small agonist 3 was similar at hH1R and gpH1R, whereas the relative affinities of the 2-phenylhistamines 12, 14, and 15 and of the histaprodifens 19 and 20 were ∼3- to 7-fold higher at gpH1R than at hH1R. These differences fit to the differences in relative agonist potencies observed in the GTPase assay (Table 2). In agreement with the GTPase studies (Table 3), the H1R antagonists triprolidine (31), arpromidine (35), and BU-E 47 (36) also all exhibited significantly higher binding affinities at gpH1R than at hH1R (Table 4).
We also studied the impact of the F153L/I433V mutation in hH1R on ligand-affinities. The double mutation significantly decreased the affinity of hH1R for histamine and 2-(2-thiazolyl)ethanamine (3) (Table 4). Similar data were obtained for the comparison of hH1R and gpH1R. Additionally, in membranes expressing hH1R-F153L/I433V, the relative affinities of 2-phenylhistamines and histaprodifens were increased relative to hH1R, but with the exception of methyl-histaprodifen (19), those changes were not as marked as for the comparison of hH1R and gpH1R. The affinities of triprolidine (31), arpromidine (35), and guanidine 36 at hH1R and hH1R-F153L/I433V were similar.
Potencies and Efficacies of H1R Agonists at hH2R and gpH2R in the GTPase Assay. The question arose whether H1R agonists, originally designed for gpH1R in comparison to gpH2R, interact differentially with the corresponding human HxRs. To address this question, we analyzed the effects of H1R agonists on GTP hydrolysis in Sf9 membranes expressing H2R-GsαS fusion proteins. We examined all H1R agonists shown in Fig. 1 and listed in Table 2 (1–23) but included only those compounds into Table 5 that actually exhibited agonistic activity at H2Rs. To account for the fact that the potency of histamine in the GTPase assay in membranes expressing H1Rs and H2Rs differs by almost 10-fold (Tables 2 and 5) (Kelley et al., 2001), we focused on the comparison of relative potencies of H1R agonists.
2-Methylhistamine (2) and 2-(2-thiazolyl)ethanamine (3) were strong partial agonists at gpH2R with moderate (2.3- to 5-fold) gpH1R/gpH2R selectivity. The introduction of a (substituted) phenyl group at position 2 of the imidazole ring greatly reduced the efficacy of H1R agonists at gpH2R and further increased gpH1R/gpH2R selectivity in terms of potency. Several 2-phenylhistamines (11, 13–15, 17, and 18) and histaprodifens 19 to 21, and 23 were devoid of agonistic activity at gpH2R-GsαS.
The analysis of histaprodifens at gpH2R-GsαS revealed the existence of a strong partial H1R agonist/moderate partial H2R agonist, Nα-(imidazolylethyl)histaprodifen (22) (Tables 2 and 5). The H2-agonistic activity of this compound can be explained by its structural similarity with guanidines 33 to 38 (Figs. 1 and 2) that are potent H2R agonists (Buschauer, 1989; Kelley et al., 2001).
Although histamine was similarly potent at stimulating GTP hydrolysis in Sf9 membranes expressing hH2R-GsαS and gpH2R-GsαS, 2-methylhistamine (2) and 2-(2-thiazolyl)ethanamine (3) were significantly less potent agonists at hH2R-GsαS than at gpH2R-GsαS and showed greater hH1R/hH2R selectivity (8.4- to 11.2-fold) than gpH1R/gpH2R selectivity (2.3- to 5-fold). If one considers the absolute EC50 values of compound 3 for GTPase activation in membranes expressing hH1R and hH2R-GsαS, the selectivity for hH1R becomes even more striking (75-versus 23-fold for gpHxRs). In contrast to compound 3, another small H1R agonist, betahistine (4), exhibited considerably higher gpH1R/gpH2R selectivity (10-fold) than hH1R/hH2R selectivity (3.5-fold). Similar to the data obtained for gpH2R, several 2-phenylhistamines (11, 13–15, 17, and 18) and histaprodifens (19–21 and 23) were devoid of agonistic activity at hH2R-GsαS. As was the case for gpHxRs, Nα-(imidazolylethyl)-histaprodifen (22) was a strong partial hH1R agonist/moderate partial hH2R agonist. There were no significant differences in the interaction of histaprodifens at hH2R-GsαS and gpH2R-GsαS. Finally, the efficacies of the 2-phenylhistamines 6 to 9 were significantly lower at hH2R-GsαS than at gpH2R-GsαS (Table 5) and therefore in the same order as observed for hH1R and gpH1R (Table 2).
Discussion
Pharmacological Differences between hH1R and gpH1R. hH1R is an important drug target for treatment of allergic diseases (second-generation H1R antagonists) and sedation (first-generation H1R antagonists) (Hill et al., 1997). Preliminary data indicate that pharmacological differences between H1R species isoforms exist (Chang et al., 1979; Seifert et al., 1994; Elz et al., 2000), but a systematic analysis of this topic has not yet been conducted. Therefore, we studied recombinant hH1R and gpH1R with 23 H1R agonists (1–23) (Fig. 1), 22 H1R antagonists (24–45) (Fig. 2), and two H2R agonists (46 and 47) under identical experimental conditions, using the GTPase assay (Fig. 5, Tables 2 and 3) and [3H]mepyramine binding assay (Tables 1 and 4) as read-out.
There were no significant differences between hH1R and gpH1R with respect to the potencies and efficacies of small agonists (1-4) in the GTPase assay (Fig. 5; Table 2). With respect to bulkier ligands, however, we found significant differences between hH1R and gpH1R. Specifically, H1R agonists of the 2-phenylhistamine class (6–17) and histaprodifen class (19–23) were generally more potent and efficacious in the GTPase assay in membranes expressing gpH1R than in membranes expressing hH1R (Fig. 5; Table 2). Additionally, in the binding assay, 2-phenylhistamines and histaprodifens exhibited higher relative affinities for gpH1R than for hH1R (Table 4). The differential interaction of 2-phenylhistamine derivatives with gpH1R and hH1R is independent of the agonist or antagonist properties of compounds (compare 17 and 40; Tables 2 and 3). High constitutive GPCR activity results in high agonist potency and efficacy (Kenakin, 1996; Seifert and Wenzel-Seifert, 2002), but we did not find differences in constitutive activity between hH1R and gpH1R studying inverse agonists. Finally, several first-generation H1R antagonists (30–32) and particularly arpromidine-type H1R antagonists (33–38) showed higher affinities for gpH1R than for hH1R. Our data concerning the affinity of ([3H])mepyramine for hH1R and gpH1R (Tables 1 and 3) fit very well to previously published data on H1R species isoforms expressed in native brain (Chang et al., 1979). Collectively, our data suggest that the ligand-binding site of gpH1R exhibits a higher conformational flexibility than the ligand-binding site of hH1R, allowing bulky compounds like 2-phenylhistamines, histaprodifens, mepyramine-type antagonists, and guanidines to dock more efficiently into gpH1R than into hH1R.
Most of the previous H1R antagonist development had been conducted with guinea pig models (Hill et al., 1997). Thus, from a therapeutic standpoint, it is fortunate that there are no or only small differences between hH1R and gpH1R with respect to commonly used first-generation H1R antagonists (e.g., 24–28, 30, and 32) and second-generation antagonists (41–45). Nevertheless, with regard to the design of H1R agonists and guanidine-type H1R antagonists, which are currently used only as experimental tools (Zingel et al., 1995; Hill et al., 1997), the H1R species isoform is of much greater relevance.
Differences in Electrophoretic Mobility between hH1R and gpH1R. A previous study showed that H1R isoforms expressed in brain from various species exhibit different migration in SDS-PAGE (Ruat and Schwartz, 1989). These data prompted us to study the electrophoretic mobility of recombinant FLAG epitope-tagged recombinant hH1R and gpH1R (Fig. 4). In agreement with the data concerning native H1R species isoforms, recombinant H1R species isoforms showed different migration in SDS-PAGE. hH1R exhibited a moderately higher molecular mass (∼76 kDa) than predicted (∼56 kDa) (Fukui et al., 1994). hH1R migrated as mixture of N-glycosylated and nonglycosylated protein, as assessed by the effect of the inhibitor of N-glycosylation, tunicamycin (Seifert and Wenzel-Seifert, 2001). Recombinant gpH1R exhibited very different migration in SDS-PAGE than hH1R, i.e., we detected faint diffuse ∼36- and ∼50-kDa bands and intense crisp ∼16- and ∼30-kDa bands in Sf9 membranes expressing gpH1R. In contrast to the results obtained with hH1R, tunicamycin had no effect on migration of gpH1R, pointing to different types of glycosylation in the two H1R species isoforms. Currently, we do not know the identity of the multiple bands in Sf9 membranes expressing gpH1R, but atypical migration of GPCRs in SDS-PAGE has been repeatedly observed (Grünewald et al., 1996; Kelley et al., 2001; Seifert and Wenzel-Seifert, 2001). Because even complex supramolecular structures such as GPCR dimers are preserved in SDS-PAGE (Fukushima et al., 1997; Hebert and Bouvier, 1998; Kelley et al., 2001), it is possible that the different electrophoretic mobilities of hH1R and gpH1R reflect different GPCR conformations. The different GPCR conformations may be associated with the specific pharmacological properties of H1R species isoforms.
Molecular Basis for the Pharmacological Differences between hH1R and gpH1R. Site-directed mutagenesis was successful at identifying the molecular basis for pharmacological differences between species isoforms of H2R and H3R (Ligneau et al., 2000; Kelley et al., 2001). We wished to apply the same strategy to H1R species isoforms. The pharmacological data discussed above indicate that the ligand-binding pocket of gpH1R is more flexible than the binding pocket of hH1R. Thus, gpH1R may possess smaller amino acid substitutions in the ligand-binding domain than hH1R so that bulkier structures are accommodated more easily in gpH1R than in hH1R. In fact, the amino acid substitutions at positions 153 (TM IV) and 433 in hH1R (TM VI) are bulkier than the corresponding amino acid substitutions in gpH1R (Phe→Leu exchange in TM IV and Ile→Val exchange in TM VI, respectively). Nevertheless, the Phe→Leu exchange in TM IV and the Ile→Val exchange in TM VI only partially explain the differences in agonist-pharmacology between hH1R and gpH1R (Tables 2 and 4). Moreover, with respect to the differences in antagonist-pharmacology, the Phe→Leu- and Ile→Val exchanges between hH1R and gpH1R are irrelevant (Tables 3 and 4). Thus, additional mutagenesis studies targeting the top portions of TM II and TM VII are required to elucidate the molecular basis for the pharmacological differences between hH1R and gpH1R (Fig. 3).
Although our mutagenesis studies were disappointing in terms of elucidating the molecular basis for the pharmacological differences between hH1R and gpH1R, our studies revealed an unexpected role of Phe-153 and Ile-433 in H1R expression and folding. Specifically, Phe-153→Leu-153- or Ile-433→Val-433 exchange in hH1R (hH1R→gpH1R) resulted in poor receptor expression, low [3H]mepyramine affinity, and functional inactivity (Table 1). Moreover, the mutations grossly altered the electrophoretic mobility of hH1R (Fig. 4). The double mutation rescued the single mutants in terms of function (Tables 1, 2, 3, 4), and it also changed electrophoretic mobility (Fig. 4). These data suggest that the couples Phe-153/Ile-433 or Leu-153/Val-433 are required for a functionally active H1R. Thus, even conservative amino acid substitutions in TM regions can have profound effects on antagonist affinity, expression, and folding of a GPCR.
Comparison of the Effects of H1R Agonists at Recombinant and Native gpH1R. Historically, the guinea pig ileum has been the standard system for the design of H1R ligands (Zingel et al., 1995; Hill et al., 1997). Therefore, it is important to compare the intact organ data with the results regarding recombinant H1R. Although many highly potent H2R and H3R agonists (i.e., ligands ∼50- to 150-fold more potent than histamine) were developed (Hill et al., 1997), the design of potent H1R agonists has been a much more difficult task. In fact, the most potent 2-phenylhistamine, 2-(3-trifluoromethylphenyl)histamine (17), is only 1.3-fold more potent, and methylhistaprodifen (20) just ∼3.5-fold more potent, than histamine in the guinea pig ileum (Leschke et al., 1995; Zingel et al., 1995; Elz et al., 2000) (Table 2).
The expression level of H1R in the guinea pig ileum is much lower than in the Sf9 cell expression system (Table 1) (Hill et al., 1997). If there had been differences in receptor reserves between the two systems, we would have expected higher agonist efficacies in the recombinant system than in the native system (Hoyer and Boddeke, 1993; Kenakin, 1996). The opposite was the case, however (Table 2) (Leschke et al., 1995; Zingel et al., 1995; Elz et al., 2000). Thus, we can rule out differences in receptor reserves accounting for the pharmacological differences between the two systems.
All agonists studied with the exception of 1, 3, 22, and 23 were more potent at the recombinant gpH1R than at the native gpH1R (Table 2). The increase in potency at the recombinant gpH1R ranged from ∼2-fold to almost 20-fold and was most pronounced for the 2-phenylhistamines 7, 11, and 13. Several explanations that are not mutually exclusive could account for the potency differences in the two systems. First, there may be substantial penetration barriers for certain agonists to reach the tunica muscularis of the ileum. Second, compounds may accumulate in certain irrelevant cells, i.e., epithelial cells, and/or, third, they may be subject to degradation. These pharmacokinetic factors are very unlikely to be of relevance when assessing the effects of ligands in membrane fragments of insect cells. Fourth, it is possible that differences in gpH1R glycosylation in insect cells versus native tissue contribute to the pharmacological differences in the two systems. Indeed, changes in glycosylation of H1R have already been shown to alter the pharmacological properties of the GPCR (Mitsuhashi and Payan, 1989). Fifth, we studied coupling of H1Rs to insect cell Gq-proteins (Houston et al., 2002), and the specific type of Gq-protein may have an impact on the pharmacological properties of gpH1R (Wenzel-Seifert and Seifert, 2000). In contrast to the above-discussed data, the high potency of compounds 22 and 23 in the guinea pig ileum does not fit to the results obtained with recombinant gpH1R. Additional studies with 22, 23, and closely related new compounds must be performed to clarify this discrepancy.
Collectively, previous studies on the guinea pig ileum resulted in considerably lower potencies of most H1R agonists than in the recombinant system. Although the high potency of H2R-and H3R agonists has not yet been achieved for H1R agonists, our present study shows that gpH1R agonists with up to ∼12-fold higher potency than histamine exist, provided that the GPCR is analyzed in the GTPase assay using membranes. Thus, future studies on the design of H1R agonists should be complemented with the recombinant system described herein.
Species Differences in Pharmacological Properties of HxRs. H2R, H3R, and H4R all exhibit species differences in their pharmacological properties (Ligneau et al., 2000; Lovenberg et al., 2000; Kelley et al., 2001; Liu et al., 2001). Thus, we were not too surprised to uncover differences in the pharmacological properties of H1R species isoforms. The species differences in pharmacological properties of HxRs extend into HxR subtype-selectivity of compounds. There are numerous efficacious H1R agonists of the 2-phenylhistamine and histaprodifen class with high gpH1R/gpH2R selectivity (Tables 2 and 5) (Leschke et al., 1995; Zingel et al., 1995; Elz et al., 2000). For the analysis of hH1R, however, one has to consider the fact that 2-phenylhistamines and histaprodifens possess substantially lower efficacies than histamine (Table 2). Unexpectedly, 2-(2-thiazolyl)ethanamine (3), a small agonist with full efficacy at hH1R, exhibited a larger hH1R/hH2R-than gpH1R/gpH2R-selectivity (Tables 2 and 5). Thus, for the analysis of hH1R with a selective hH1R agonist, compound 3 may be the ligand of choice. These findings emphasize the importance to study hHxR isoforms for the development of hHxR ligands. Future studies will have to answer the question whether the species differences in pharmacological properties of HxRs reflect species-specific adaptations to as yet unidentified endogenous and/or exogenous HxR ligands.
Although H1Rs and H2Rs are structurally quite distinct from each other (only ∼40% homology) (Traiffort et al., 1994; Hill et al., 1997), there is a common aspect in the pharmacological properties of these GPCRs, i.e., the preferential interaction of bulky agonists with gpHxRs relative to hHxRs. Most notably, arpromidine-derived guanidines represent a class of ligands that exhibit higher affinities for gpH1R and gpH2R relative to hHxRs (Tables 3 and 4) (Kelley et al., 2001). Those differences may indicate that gpHxRs in general possess a higher conformational flexibility than hHxRs.
Acknowledgments
We express gratitude to Dr. G. Lambrecht (Department of Pharmacology, University of Frankfurt/Main, Germany) for providing dimethindene enantiomers, to Dr. H. Fukui (Department of Pharmacology, University of Tokushima, Japan) for providing cDNAs for hH1R and gpH1R, and Dr. E. Ross (Department of Pharmacology, University of Southwestern Medical Center, Dallas, TX) for donating baculoviruses encoding RGS proteins. We also thank Dr. F. Schalkhausser (Institute of Pharmacy, University of Regensburg, Germany) for the synthesis of guanidine 39, Dr. K. Kramer (Institute of Pharmacy, Free University of Berlin, Germany) for the synthesis of compounds 2, 3, 22, 23, and 40, Dr. W. Toegel (Institute of Pharmacy, Free University of Berlin, Germany) for the preparation of compound 18, and C. Houston (Department of Pharmacology and Toxicology, University of Kansas, Lawrence, KS) for help with the cell culture. Thanks are also due to the reviewers for their constructive critique.
Footnotes
-
This work was supported by the National Institutes of Health COBRE Award 1 P20 RR15563 and matching support from the State of Kansas and the University of Kansas (R.S.), a grant from the Army Research Office (DAAD 19-00-1-006) (R.S.), grants of the Deutsche Forschungsgemeinschaft (W.S. and A.B.), Fonds der Chemischen Industrie (T.B., W.S., A.B., and S.E.), the Bundesministerium für Bildung und Forschung (T.B.), and the Deutscher Akademischer Austauschdienst within the International Quality Network “Medicinal Chemistry” (S.E. and A.B.).
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.103.049619.
-
ABBREVIATIONS: HxR, histamine H1-, H2-, H3-, or H4-receptor; h, human; gp, guinea pig; GPCR, G-protein-coupled receptor; TM, transmembrane domain; RGS protein, regulator of G-protein signaling; GAIP, Gα-interacting protein; PAGE, polyacrylamide gel electrophoresis; gpH2R-GsαS, fusion protein of the guinea pig histamine H2-receptor and the short splice variant of Gsα;hH1R, human histamine H1-receptor; hH2R, human histamine H2-receptor; hH2R-GsαS, fusion protein of the human histamine H2-receptor and the short splice variant of Gsα; hH1R-F153L, human histamine H1-receptor bearing a Phe→Leu exchange at position 153; hH1R-I433V, human histamine H1-receptor bearing an Ile→Val exchange at position 433; hH1R-F153L/I433V, human histamine H1-receptor bearing a Phe→Leu exchange at position 153 and an Ile→Val exchange at position 433.
- Received January 28, 2003.
- Accepted February 21, 2003.
- The American Society for Pharmacology and Experimental Therapeutics