Distinct pharmacology of rat and human histamine H₃ receptors: role of two amino acids in the third transmembrane domain

Abstract

Starting from the sequence of the human histamine H₃ receptor (hH₃R) cDNA, we have cloned the corresponding rat cDNA. Whereas the two deduced proteins show 93.5% overall homology and differ only by five amino acid residues at the level of the transmembrane domains (TMs), some ligands displayed distinct affinities. Thioperamide and ciproxifan were about 10 fold more potent at the rat than at the human receptor, whereas FUB 349 displayed a reverse preference. Histamine, (R)α-methylhistamine, proxyfan or clobenpropit were nearly equipotent at H₃ receptors of both species. The inverse discrimination patterns of ciproxifan and FUB 349 were partially changed by mutation of one amino acid (V122A), and fully abolished by mutation of two amino acids (A119T and V122A), in TM3 of the rH₃R located in the vicinity of Asp¹¹⁴ purported to salt-link the ammonium group of histamine. Therefore, these two residues appear to be responsible for the distinct pharmacology of the H₃R in the two species.

Introduction

Whereas the histamine H₃ receptor (H₃R) was initially identified in the rat brain (Arrang et al., 1983; 1987), its presence in the human brain was confirmed a few years later (Arrang et al., 1988). In both cases a functional test, the inhibition of [³H]-histamine release from depolarized brain slices, was used, but the pharmacological characterization of the human H₃R has remained preliminary since the availability of fresh brain tissues obtained during neurosurgery is limited. Nevertheless there were some indications that the pharmacology of the human and the rat H₃R may slightly differ (Arrang et al., 1988; West et al., 1999 and X. Ligneau, unpublished observation).

With the recent cloning of the human H₃R (hH₃R) (Lovenberg et al., 1999), it became feasible to determine with greater precision the apparent affinity of ligands at this receptor and assess the existence of species differences. Namely, for this purpose we have cloned the rat H₃R (rH₃R) starting from the published sequence of the hH₃R and established permanent cell lines expressing the hH₃R or rH₃R. This allowed us to identify ligands displaying distinct apparent affinities at the H₃R of the two species. Then we tried to identify the amino acid residues responsible for such discrimination using site-directed mutagenesis of the rH₃R.

Methods

Cloning of the rH₃R cDNA

A rat striatal cDNA library (4×10⁶ phages; Stratagene, St-Quentin-en-Yvelines, France) was screened at high stringency with a ³²P-labelled fragment (607 bp), obtained by RT–PCR amplification of total mRNAs from rat cerebral cortex using primers based on the sequence of the third transmembrane domain (TM3 nucleotides 299–331) and the third intracellular loop (nucleotides 1601–1637) of the hH₃R, respectively (Lovenberg et al., 1999). Bluescript KS(+) plasmids were recovered from sixty positive clones and their cDNA inserts sequenced. Some of them exhibited a full-length open reading frame encoding a 445-amino acid protein corresponding to the rH₃R.

Cloning of the hH₃R cDNA

Screening of a human striatum cDNA library (Stratagene) with the same probe led to the isolation of five positive clones. Among them, one exhibited a full-length cDNA sequence displaying a 100% identity with the hH₃R cDNA recently described (Lovenberg et al., 1999).

Stable transfection of CHO-K1 cells

cDNA inserts corresponding to the full-length coding sequences of the rH₃R and hH₃R, were ligated into the mammalian expression vector pCIneo (Promega, Charbonnières, France). CHO-K1 cells were transfected using SuperFect (Qiagen, Courtaboeuf, France). Stable transfectants were selected with 2 mg ml⁻¹ of G418 and tested for [¹²⁵I]-iodoproxyfan binding (Ligneau et al., 1994). Two clones expressing ∼300 fmol mg⁻¹ protein of rH₃R and hH₃R binding sites were selected for further characterization and maintained in the presence of 1 mg ml⁻¹ of G418.

Site directed mutagenesis of the rH₃R

Rat H₃-receptor mutants were constructed using the Transformer Site-Directed Mutagenesis kit (Clontech, St-Quentin-en-Yvelines, France). The rH₃R cDNA subcloned into pCIneo plasmid (0.2 μg) was used for synthesis of mutant cDNA strands using T4 DNA polymerase (4 units; Clontech) and T4 DNA ligase (6 units; Clontech), in the presence of the mutagenic and selection primers phosphorylated with T4 polynucleotide kinase (Roche, Meylan, France). The mutagenic primer (5′-cctactgtgtacctcctcggccttcaacatc-3′) was designed to introduce the required mutations within the third transmembrane domain of the rH₃R and to suppress a BbsI restriction site which was found in the wild-type rH₃R sequence. The selection primer (5′-cgagacagaaaaaacacttgcgtttctgata-3′) was designed to suppress a BbsI restriction site found within the pCIneo plasmid sequence. Following selection by BbsI restriction endonuclease digestion, the mutant plasmids were amplified in competent BMH71-18mutS cells (Clontech) and further isolated after complete BbsI digestion of parental plasmids (40 units; 2 h). Following amplification in Top10 cells (Invitrogen, Groningen, The Netherlands), undigested (mutant) plasmids were prepared and sequenced using Taq FS DNA polymerase (Perkin-Elmer, Courtaboeuf, France).

Transient expression in Cos-1 cells

Cos-1 cells were grown to 40% confluence in DMEM-Ham F12 medium (Life Technologies, Cergy-Pontoise, France) with 10% foetal calf serum (Valbiotech, Paris, France) in 96 mm Petri dishes and transfected with 5 μg of the plasmid pCIneo-rH₃R, pCIneo-[122A]rH₃R, pCIneo-[119T/122A]rH₃R and pCIneo-hH₃R, using 30 μl of SuperFect Transfection Reagent (Qiagen). Two days later, the cells were harvested and membranes prepared for binding assays.

[¹²⁵I]-Iodoproxyfan binding assays

Transfected CHO or Cos-1 cells were washed and homogenized with a polytron in ice-cold binding buffer (Na₂HPO₄/KH₂PO₄ 50 mM, pH 6.8) and assays performed as described (Ligneau et al., 1994). Briefly, aliquots of membrane suspension (5–15 μg of protein) were incubated for 60 min at 25°C with 25 pM [¹²⁵I]-iodoproxyfan alone, or together with competing drugs (200 μl, final volume). The nonspecific binding was determined using imetit (1μM).

Analysis of data

IC₅₀ values were determined using an iterative least-squares method derived from that of Parker & Waud (1971). K_i values of ligands were calculated from their IC₅₀ values, assuming a competitive antagonism and by using the relationship (Cheng & Prussoff, 1973):

S and K_D represent the concentration and dissociation constant of the radioligand respectively.

Results

Comparison of the rH₃R and hH₃R amino acid sequences

Sequence analysis of the rH₃R cDNA revealed a full-length open reading frame encoding a protein of 445 amino acids displaying 93.5% overall homology with the human receptor. Amino acid sequence alignments of the two cloned receptors showed only five different residues within the seven transmembrane domains (TMs). Among the latter, two were found in TM3, close to an aspartic acid residue known to be conserved in all aminergic receptors, and corresponded to A119T and V122A transitions (Figure 1).

Figure 1.

Putative membrane topology of the histamine H₃ receptor. The amino acid sequence of the third transmembrane domain of the rat (rH₃R) and the human (hH₃R) receptor is shown. The open box indicates the position of aspartic acid 114, known to be conserved in all aminergic receptors. The grey boxes indicate the amino acids in position 119 and 122. Mutations at these position in the rat receptor are underlined.

Comparison of the rH₃R and hH₃R pharmacological profiles in CHO cells

Specific [¹²⁵I]-iodoproxyfan binding to membranes of CHO(rH₃R) and CHO(hH₃R) was monophasic and saturable (B_max ∼300 fmol mg⁻¹ protein). Computer analysis by nonlinear regression using a one-site cooperative model led to K_D values of 68±15 pM and 50±7 pM at rH₃R and hH₃R, respectively. [¹²⁵I]-Iodoproxyfan binding was inhibited by a series of H₃-receptor ligands with K_i values defining a pharmacological profile of the rH₃R distinct from that of the hH₃R (Table 1). Histamine and the agonist (R)α-methylhistamine displaced specific binding (n_H=0.7–0.8) with similar affinities at the two receptors. The pseudo Hill coefficient of the antagonists did not significantly differ from unity. Among the latter, proxyfan and clobenpropit displayed a similar affinity at the two receptors, whereas thioperamide and ciproxifan were about 10-fold more potent at the rH₃R and, in contrast, FUB 349 was about 6 fold more potent at the hH₃R (Table 1).

Table 1.

Compared potencies of H₃-receptor ligands on inhibition of [¹²⁵I]-iodoproxyfan binding to rH₃R and hH₃R stably expressed in CHO cells

Pharmacological analysis of mutant rat receptors in Cos-1 cells

Sequence analysis of mutant plasmids obtained by site-directed mutagenesis revealed the two expected mutations of the rat receptor, i.e., mutation of alanine 119 to threonine and mutation of valine 122 to alanine ([119T, 122A] rH₃R). However, one set of plasmids contained only the mutation of valine 122 ([122A] rH₃R) (Figure 1). [¹²⁵I]-Iodoproxyfan binding to membranes of Cos-1 cells expressing rH₃R, [122A] rH₃R, [119T, 122A] rH₃R and hH₃R occurred with similar K_D values, i.e. 71±4; 57±13; 45±5 and 41±6 pM, respectively. As with CHO cells, ciproxifan inhibited H₃R binding to Cos-1 cell membranes with higher potency at the wild-type rat, as compared to the human receptor (13 fold difference). This difference was reduced with the single and abolished with the double mutation) (Figure 2).

Figure 2.

Inhibition of [¹²⁵I]-iodoproxyfan binding to mutant rat receptors by ciproxifan and FUB 349. Membranes of Cos-1 cells expressing wild-type rat receptors (rH₃R), mutant [122A] rat receptors, mutant [119T, 122A] rat receptors or wild-type human receptors (hH₃R) were incubated with 25 pM [¹²⁵I]-iodoproxyfan and ciproxifan or FUB 349 in increasing concentrations. Each point represents the mean value from two different experiments with triplicate determinations each. The K_i values (nM) of ciproxifan and FUB 349 obtained for each receptor are indicated.

The opposite difference in potency of FUB 349 in wild-type receptors detected in CHO cells was confirmed in Cos-1 cells (∼5 fold higher potency at hH₃R than at rH₃R) and, this time, the mutations resulted in enhanced potency (over 10 fold in the double mutation) (Figure 2).

In contrast, the potency of clobenpropit was similar at the H₃R of the two species expressed in Cos-1 cells, i.e., 1.5–2 nM (not shown).

Discussion

Using recombinant receptors expressed in different cell lines, the present work first confirms that the H₃Rs can be differentiated pharmacologically in two species and, then, identifies the area very likely responsible for this difference.

Previous indications of such species differences, mainly between the hH₃R and rH₃R were derived from either functional or binding assays performed with fresh brain tissues. In agreement, the prototypical H₃R antagonist thioperamide was found to be slightly (4 fold) less potent at the H₃R modulating [³H]-histamine release from depolarized human when compared to rat brain slices, K_i values being 16 nM (Arrang et al., 1988) and 4 nM (Arrang et al., 1987), respectively. Other functional or binding studies performed with various ligands led to even higher K_i values for this compound at the hH₃R, i.e., 85–200 nM (Cherifi et al., 1992; West et al., 1999). In the same way the antagonist ciproxifan displayed significantly higher potency at the rH₃R when compared to the hH₃R in fresh brain tissues using either functional or binding assays performed under parallel conditions (Ligneau et al., 1998 and X. Ligneau, unpublished observations). Interestingly, however, both histamine and (R)α-methylhistamine were found to be nearly equipotent at the native rH₃R and hH₃R in brain (Arrang et al., 1987, 1988; West et al., 1999).

Here we come to similar conclusions, i.e., thioperamide and ciproxifan are about 10 fold more potent at the rH₃R than at the hH₃R, whereas histamine, (R)α-methylhistamine and the two antagonists clobenpropit (Van der Goot et al., 1992) and proxyfan (Stark et al., 1998a) were nearly equipotent.

Moreover, we identified one compound, the antagonist FUB 349 (Stark et al., 1998b), displaying a reverse preference, i.e., being about 5 fold more potent at the human than at the rat receptor.

While this work was in progress, Lovenberg et al. (2000) have also cloned a rH₃R with a sequence 100 per cent identical to that we have established independently, and confirmed the higher potency of thioperamide (K_i=4 nM) at this receptor when compared to its human counterpart (K_i=58 nM or K_i=20 nM in Lovenberg et al., 1999).

Although these differences in potency of some H₃R ligands in the two species are of rather limited amplitude, they are reproducibly found with native as well as recombinant receptors expressed in various cell lines. This leaves little doubt about the existence of a pharmacological heterogeneity.

The identification of residues responsible for this heterogeneity was facilitated by the realization that the rH₃R and hH₃R sequences differed by only five amino acid residues, at the level of the putative TM helices where ligands are thought to bind. Among these helices, TM3 was a good candidate since, at this level, the rH₃R differs from the hH₃R by two residues located in vicinity to the aspartate residue (Asp¹¹⁴) present in all aminergic receptors and purported to salt-link the ammonium group of histamine and agonists. Mutation of these two residues, i.e., Ala¹¹⁹ into Thr¹¹⁹ and Val¹²² into Ala¹²², to obtain a partially ‘humanized' rat H₃ receptor led to the expected changes. In agreement: (i) the affinity of ligands not discriminating hH₃R and rH₃R, e.g., [¹²⁵I]-iodoproxyfan or clobenpropit was not significantly modified, and (ii) in contrast the affinity of a rH₃R-preferring ligand, ciproxifan, was reduced, whereas that of FUB 349, a hH₃R-preferring ligand, was enhanced so that the affinity of these compounds did not differ anymore from corresponding values at the hH₃R (Figure 2).

The mutated amino acids may modify, e.g., hydrophobic interactions, which are presumed to have greater influence on the binding of lipophilic antagonists than on that of hydrophilic agonists. The purported salt link of basic compounds like histamine, (R)α-methylhistamine, and clobenpropit with Asp¹¹⁴ seems to be of greater importance for these compounds than their hydrophobic interactions. Such hydrophobic interactions may involve the mutated transmembrane domain and the relative potencies at rat and human receptors be determined by the relative positioning of the imidazole ring, the phenyl (or other hydrocarbon residue) and the polar group, if present in the antagonist.

Interestingly, the guinea-pig H₃R slightly differs from both hH₃R and wild-type or mutated rH₃R by the presence of threonine and valine residues at corresponding 119 and 122 positions respectively (Tardivel-Lacombe et al., 2000). Since the isolated guinea-pig ileum is currently used to evaluate the potency of H₃R ligands (Hill et al., 1997), it would be interesting to assess the role of the two critical TM3 amino acid residues in differences in H₃R pharmacology in this species, if any.

In addition, more extensive structure-activity together with modelling studies are likely to provide more details about a possible interaction of ligands with the receptor at this level and should facilitate the rational design of novel ligands to be used as drugs in humans.

Abbreviations

H₃R

histamine H₃ receptor

hH₃R

human histamine H₃ receptor

rH₃R

rat histamine H₃ receptor

TM

transmembrane domain