Abstract
Urotensin II (U-II; cyclo5-10[H-Glu-Thr-Pro-Asp-Cys-Phe-Trp-Lys-Tyr-Cys-Val-OH]) is a potent vasoconstrictor in mammals, and it is postulated that it plays a central role in cardiovascular homeostasis. Thus, we initiated a structure-to-function analysis of this peptide characterized by a N-terminal tail and a cyclic core formed through a disulfide bridging. A total of 41 analogs focusing on these characteristics were developed and evaluated using a binding assay on membranes from a stable HEK-293 cell line containing the human or rat U-II receptor, a functional assay for Ca2+ mobilization on transiently transfected CHO-K1 cells with the human or rat U-II receptor, and a rat thoracic aorta bioassay. At first, the focus was applied on peptide compounds containing exocyclic modifications. From this series, it appeared that only valine-11 played a significant role although it is not an essential amino acid. Similarly, endocyclic and ring transformations of hU-II were also studied. In most cases, a detrimental effect on affinity and biological activity was observed. However, two compounds, [Tyr6]hU-II and [Phe9]hU-II, retained affinity and activity. So far, our binding, functional, and pharmacological data clearly demonstrated the minor contribution of the N-terminal segment and the essential role of the cyclic structure. More particularly, three residues within the loop, i.e., Trp-7, Lys-8, and Tyr-9, are required for receptor recognition and activation. This three-pole feature, kept by the disulfide bond in a correct spatial arrangement, appears as the key pharmacophore for the U-II receptor.
Urotensin II (U-II), a potent mammalian vasoconstrictor, was first isolated and characterized from the Gillichthys mirabilis (goby) urophysis, a neurohemal organ (Pearson et al., 1980). Recently, human cDNA was cloned, and hU-II appeared as an 11 amino acid peptide with cysteine residues in positions 5 and 10, forming a bridge through their side chains (Coulouarn et al., 1998). A comparison of the U-II primary structures revealed a striking homology among species. From fish to humans, evolution favored the conservation of residues 5 to 10, whereas the N-terminal segment varied in length and composition. In addition, position 4 is occupied by the acidic residue Asp or Glu, whereas position 11 is characterized by the presence of the hydrophobic residue Val or Ile.
The most striking effects of U-II are observed in the cardiovascular arterial system in which it is frequently identified as the most potent vasoconstrictor ever described. Using tissues from several species, however, Douglas et al. (2000a) carried out a study on the vasoconstrictor activity of human U-II (hU-II), and they showed that both potency and efficacy vary significantly between species, individuals, and vessels. As such, hU-II was described as the most potent vasoconstricting compound on the rat thoracic aorta, but in mouse, whatever the vessels tested, no vasoconstriction was recorded. On the other hand, hU-II was a potent vasoconstrictor of nonhuman primate arteries (Ames et al., 1999; Douglas et al., 2000b). In absence of endothelium, radial, coronary, and mammary human arteries demonstrated a moderate, yet inconsistent, constriction to hU-II (Maguire et al., 2000; Paysant et al., 2001). Removal of endothelium from specific human arteries appeared to unmask the contractile respond to hU-II (Hillier et al., 2001). In contrast, the abdominal and small pulmonary arteries showed a potent vasodilation to hU-II in presence of endothelium (Stirrat et al., 2001). Additional in vitro studies have shown that hU-II is a potent inotropic agent on human cardiac muscle (Russell et al., 2001) and that it induces vascular smooth muscle cell proliferation (Watanabe et al., 2001). In vivo reports regarding U-II vascular effects appeared inconclusive since two independent studies using identical procedure showed opposite results (Böhm and Pernow, 2002; Wilkinson et al., 2002). Nevertheless, the potency of hU-II as a vasoconstricting, vasodilating, or inotropic agent strongly suggests a key role of this peptide in the cardiovascular homeostasis.
These observations are well matched with the hypothesis that U-II might be involved in various diseases, including cardiovascular pathologies such as hypertension and arteriosclerosis. Thus, the U-II biological system exhibits a remarkable potential for the development of novel therapeutic strategies, especially those related to the treatment of cardiovascular diseases. Such developments require a precise knowledge of the pharmacophoric elements within U-II essential for the affinity and the activity of this peptide. So far, because of its resemblance to somatostatin (SST), it is conceivable that some of the critical structural features already described for SST are shared with U-II. Accordingly, both peptides contain the tripeptide Phe, Trp, Lys followed by the hydroxylated residue Thr or Tyr. Moreover, it was demonstrated in SST that an interaction between the aromatic moieties of residues Phe-6 and Phe-11 stabilizes the orientation of residues Phe-7, Trp-8, Lys-9, and Thr-10. Similarly, the disulfide bridge in U-II would act in a similar manner. Structure-activity relationship (SAR) studies of SST already showed that these pharmacophoric features are essential for the SST biological activity (Janecka et al., 2001). Furthermore, previous fish U-II (Itoh et al., 1987) and human U-II (Kinney et al., 2001, 2002; Flohr et al., 2002) SAR studies showed that the highly conserved C-terminal segment CFWKYCV is the minimal sequence required to maintain a high potency and that Lys-8 appeared as an important residue for U-II activity (Carmada et al., 2002). In addition, substitution of the U-II disulfide bridge with lactam bridges of various lengths showed that the biological activity of U-II is dependent upon the size of the cyclic structure (Grieco et al., 2002). Therefore, we initiated a SAR study of the human and rat urotensin II by targeting the essential structural features already pinpointed in the somatostatin and fish urotensin II molecules. We report the evaluation of 41 human or rat urotensin II-derived peptides, using two different functional assays on HEK293 cells stably expressing the human and rat U-II receptor: 1) mobilization of intracellular calcium and 2) constriction of the rat thoracic aorta. In addition, all peptides were tested in binding assays on membranes prepared from HEK293 cells stably expressing the rat and human U-II receptor for their ability to compete with [125I]U-II binding.
Materials and Methods
Reagents and Solvents. The following fluorenylmethyloxycarbonyl-protected amino acids were purchased from Chem-Impex International (Wood Dale, IL): Ala, Cys(Trt), His(Trt), Phe, Trp, Lys-(Boc), Tyr(tBu), Asp(OtBu), Glu(OtBu), Pro, Gln(Trt), Thr(tBu), Arg(Pbf), Orn(Boc), HomoCys(Trt), Cys(Acm), and d-Trp. Biograde TFA was obtained from Halocarbon (River Edge, NJ). Diisopropylethylamine was from Aldrich Chemical Company (Milwaukee, WI). Wang resin and benzotriazol-1-yl-oxy-tris(dimethylamino)-phosphonium hexafluorophosphate were purchased from Albatross Chemical (Montreal, QC).
Basal ISCOVE medium and fetal calf serum (FCS) were from Biochrom (Berlin, Germany). The GC-melt PCR-kit as well as the human and rat genomic DNA were purchased from CLONTECH (Palo Alto, CA). The pEAK8 mammalian episomal expression vector and the selection marker puromycin were from Edge Biosystems (Gaithersburg, MD). Dulbecco's modified Eagle's medium (DMEM), l-glutamine, HEPES, LipofectAMINE reagent, and penicillin-streptomycin were from Invitrogen (Carlsbad, CA). The pCDNA3.1(+) mammalian expression vector was from Invitrogen. The calcium-sensitive fluorescence dye Fluo-4 and Pluronic F-127 were obtained from Molecular Probes (Eugene, OR). Flashplates PLUS and monoiodinated human [125I-Tyr9]urotensin II for radioligand binding assays were from PerkinElmer Life Sciences (Boston, MA). Gentamicine, the transfection reagent FuGene 6, and Complete protease inhibitor were purchased from Roche (Basel, Switzerland). Bacitracin, EDTA-disodium salt, probenecid, MgCl2, NaCl, and sucrose were obtained from Sigma-Aldrich (St. Louis, MO).
Peptide Synthesis and Cleavage. The peptides hU-II(5-10), hU-II(2-11), hU-II(3-11), hU-II(4 -11), hU-II(1-10), [Ala5]hU-II, [d-Phe6]hU-II, [d-Lys8]hU-II, [d-Tyr9]hU-II, [des-Phe6]hU-II, [des-Trp7]hU-II, [des-Lys8]hU-II, [des-Tyr9]hU-II, [des-Lys8,Val11]hU-II, and [des-Tyr9,Val11]hU-II were custom designed and purchased from EMC Microcollections (Tübingen, Germany). All the other analogs were synthesized using a solid-phase procedure based on a fluorenylmethyloxycarbonyl chemistry/benzotriazol-1-yl-oxy-tris(dimethylamino)-phosphonium hexafluorophosphate coupling strategy designed for a home-made manual multireactor synthesizer. Wang resin was used as the solid support, and each coupling step was monitored using a ninhydrin test. Cleavage from the resin was achieved with a mixture of TFA/ethanedithiol/triisopropylsilane/phenol/water (50 ml/g 80:5:5:5:5 v/v) for 2 h. After TFA evaporation, peptides were precipitated and washed using diethyl ether. Crude peptides were dried and kept at -20°C until purification.
Peptide Purification and Characterization. About 700 mg of the crude material was dissolved in water before being injected onto a reverse-phase HPLC Jupiter C18 (15 μm; 300Å) column (250 × 21.20 mm) (Phenomenex, Torrance, CA). The purification step was carried out using a Waters Prep 590 pump system connected to a Waters model 441 absorbance detector (Waters, Milford, MA). The flow rate was fixed at 40 ml/min, and the peptide was eluted with a solvent gradient of 0 to 40% solvent B in which solvent A is aqueous NH4OH 0.05% and solvent B is CH3CN 40% in NH4OH 0.05%. Fractions homogeneity was evaluated using analytical reverse-phase HPLC with a Jupiter C18 (5 μm; 300 Å) column (250 × 4.60 mm) (Phenomenex) connected to a Beckman 128 solvent module coupled to a Beckman 168 PDA detector (Beckman Coulter, Inc., Fullerton, CA). The flow rate was at 1.0 ml/min, and the elution of the peptide was carried out with a linear gradient of 20 to 60% B in which A was TFA 0.06% and B was CH3CN 40% in aqueous TFA 0.06%. Aliquots of 20 μl were injected and analyzed. Homogeneous fractions were pooled, lyophilized, and then analyzed by matrix-assisted laser desorption and ionization-time-of-flight mass spectrometry (Voyager DE spectrometer; Applied Biosystems, Foster City, CA). The laser was set at 337 nm, and an acceleration voltage of 25 kV was applied. The compound α-cyano-4-hydroxycinnamic acid was used as a matrix for peptide inclusion and ionization. The purified peptide precursor with the right mass was cyclized using K3Fe(CN)6 as an oxidant. Thus, each peptide was dissolved in water at a concentration of 0.5 mg/ml, and 20 equivalents of oxidant was added. After 1 h, the reaction mixture was injected onto a preparative column and, as described above, purified using HPLC. The final pure peptides were characterized by analytical HPLC and matrix-assisted laser desorption and ionization-time-of-flight mass spectrometry.
Biological Activity Study. All peptides were tested pharmacologically using the rat thoracic aorta preparation. Sprague-Dawley male rats from Charles River (St-Constant, QC, Canada), weighing 250 -300g, were anesthetized with Somnotol (pentobarbital-MTC Pharmaceuticals, Cambridge, ON, Canada) and killed by exsanguination. The thoracic aorta was quickly removed and placed in oxygenated Krebs-Henseleit buffer (118 mM NaCl, 25 mM NaHCO3, 4.7 mM KCl, 1.18 mM KH2PO4, 2.5 mM CaCl2, and 22.2 mM glucose, pH 7.2) at 37°C. The endothelium was not removed, as preliminary studies showed that no differences in the data were recorded with endothelium-denuded tissues. Aorta were cleaned of their connective tissues and cut into rings (six) of ∼4 mm. Rings were suspended in 5 ml oxygenated (95% O2 and 5% CO2) bioassay vessels and maintained at 37°C. The tissues were allowed to equilibrate for 1 h under a resting tension of 1 g before being exposed to 60 mM KCl for normalization. Mechanical responses were recorded isometrically on a Grass 7D polygraph (Grass Instruments, Quincy, MA) equipped with force-displacement transducers. All peptides were dissolved in water. Concentrations of 1.0 × 10-10 to 1.0 × 10-5 M of hU-II or analogs were used to establish the concentration-response curves. All results were normalized as a percentage of the contraction obtained with KCl 60 mM. The median effective concentrations (EC50) are expressed as the mean ± S.E.M., and the n values refer to the number of rats. Differences between data were tested for significance using the Student's t test for unpaired samples. Values of p < 0.05 were considered as significant.
Cloning of Human and Rat Urotensin II Receptor. As the putative human urotensin II receptor sequence is intronless, we cloned this protein from human genomic DNA via PCR. PCR conditions, established to amplify the human GPR14 sequence were 94°C, 10 min followed by 35 cycles of 94°C, 1 min, 60°C, 1 min, 72°C, and 2 min using the GC-melt kit. Primers designed to amplify the coding sequence contained a BamHI site in the forward and a XbaI-site in the reverse primer, respectively. The urotensin receptor coding region, flanked by BamHI/XbaI sites, was cloned into the pCDNA3.1(+) mammalian expression vector and sequenced in both directions. For generation of stable cell lines, the human U-II receptor coding sequence flanked by a 5′ EcoRI site and a 3′ EcoRV site was cloned into the mammalian episomal expression vector pEAK8.
The rat U-II receptor coding sequence flanked by a 5′ EcoRI and 3′ NotI site was amplified via PCR from rat kidney cDNA and cloned into the mammalian pEAK8 expression vector. Sequences of all urotensin II receptor expressing plasmids were verified by dideoxy sequencing in both directions.
Cell Culture and Transfection. CHO K1-cells were grown in basal ISCOVE medium supplemented with 10% FCS, 2 mM l-glutamine, penicillin-streptomycin (10,000 IU/ml-10,000 μg/ml), and 25 mg/ml gentamicin at 37°C in a humidified 5% CO2 incubator. Cells were transiently transfected with the U-II-receptor cDNAs using the LipofectAMINE reagent according to the manufacturer's protocol. After 18 to 24 h following the transfection, cells were split into black-wall 96-well plates at a density of 50,000 cells/well and cultured for an additional 18- to 24-h period before being used in the functional fluorescence imaging plate reader (FLIPR) assay (described below in details) measuring intracellular Ca2+ release upon receptor activation.
FLIPR Assay. Cells were loaded in 96-well plates for 1 h (37°C, 5% CO2) with 100 μl of PBS (without Ca2+, Mg2+ and NaHCO3) containing 4 μM of the fluorescent calcium indicator Fluo-4, 0.22% Pluronic F-127 in dimethyl sulfoxide, 2.5 mM probenecid, 1 mM EGTA, and 1% FCS. Cells were then washed three times with PBS (without Ca2+,Mg2+, and NaHCO3) containing 1 mM EDTA, 0.5 mM MgCl2, and 2.5 mM probenecid. After the final wash, a 100-μl residual volume remained on the cells. Peptides were aliquoted as 2× solutions in 96-well plates and transferred by the instrument from the ligand plate to the cell plate. Fluorescence was recorded with the fluorometric imaging plate reader FLIPR (Molecular Devices, Sunnyvale, CA) over a period of 3 min. Fluorescence was recorded simultaneously in all wells in 3-s intervals during the 1st minute and in 10-s intervals during the last 2 min. Fluorescence data were generated in duplicate and repeated for at least three times. Data were analyzed by nonlinear curve fitting using GraphPad Prism version 3.0 (GraphPad Software, Inc., San Diego, CA).
Generation of Stable Human and Rat U-II Receptor Expressing Cell Lines. For the generation of stable cell lines expressing the human and rat urotensin II receptor, HEK293 cells were transfected with human- and rat-pEAK8 constructs using the FuGene 6 transfection reagent according to the supplier's protocol. Two days after transfection, cells were selected in DMEM, supplemented with 10% fetal calf serum, 20 mM HEPES, penicillin-streptomycin (10,000 IU/ml-10,000 μg/ml), and 1 μg/ml puromycin for a period of 4 weeks (37°C, 5% CO2, 95% relative humidity). Functional activity of the urotensin II receptor-expressing cell population was verified with a FLIPR assay recording urotensin II-mediated intracellular Ca2+ release, as described above.
Membrane Preparation and Radioligand Binding Assays. HEK293 cells stably expressing human or rat urotensin II receptor were cultured up to 80% confluency in DMEM, supplemented with 10% fetal calf serum, 20 mM HEPES, penicillin-streptomycin (10,000 IU/ml-10,000 μg/ml), and 1 μg/ml puromycin (37°C, 5% CO2, 95% relative humidity). Cells were washed once with ice-cold PBS and a second time with PBS containing the protease inhibitor cocktail (Complete). Cells were scraped off and centrifuged gently. The pellet was resuspended in a buffer containing 5 mM HEPES, 1 mM EDTA-disodium salt, and the cocktail of protease inhibitor Complete and then incubated on ice for 15 min. Cells were pelleted again and resuspended with a homogenizer (Unit F8B, Constant cell disruption systems; Honiley, Warwickshire, UK). The supernatant and resuspended pellet were combined and centrifuged at 50,000g (Beckman Avanti J251). The cell membrane pellet was resuspended in a buffer consisting of 20 mM HEPES, 1 mM EDTA-disodium salt, 150 mM NaCl, and 10% sucrose. Membrane aliquots were stored at -80°C. One day before the binding assay, membranes were thawed, pelleted, and resuspended in the assay buffer consisting of 20 mM HEPES, 150 mM NaCl, 1 mM EDTA-disodium salt, 160 μg/ml bacitracin, and Complete protease inhibitor (two tablets per 100 ml).
Membranes were distributed into 96-well wheatgerm-agglutinin Flashplates PLUS, incubated overnight for adsorption, and then the Flashplates PLUS were washed twice with the assay buffer. For equilibrium binding assays, 0.2 nM human [125I-Tyr9]urotensin II (initial specific activity 2200 Ci/mmol) was incubated with the indicated amounts of unlabeled competitors during 4 h and radioactivity counted in a 1450 Microbeta Wallac Jet (Wallac, Turku, Finland). Nonspecific binding was determined in the presence of 10 μM human U-II and corresponded to 31 ± 2% (n = 5, mean ± S.E.M.) of the total binding for membranes expressing the human urotensin II receptor and to 8 ± 1% (n = 5, mean ± S.E.M.) of the total binding for membranes expressing the rat urotensin II receptor. Data analysis was performed with GraphPad Prism 3.0.
Results
Pharmacological Studies (Rat Thoracic Aorta Bioassay). A series of 41 analogs (Table 1) was developed and evaluated for their biological activity, using the rat thoracic aorta paradigm (Fig. 1 for representative curves). The focus was first applied on peptide compounds containing exocyclic modifications (Table 2). Thus, the fragments obtained by shortening the N-terminal segment of hU-II, such as hUII(2-11) (compound 3), hU-II(3-11) (compound 4), hU-II(4-11) (compound 5), hU-II(5-11) (compound 6), and hU-II(5-10) (compound 2) appeared as partial agonists with EC50 values of 3.3 × 10-9, 2.8 × 10-10, 3.0 × 10-8, 6.8 × 10-7, and 1.5 × 10-6 M, respectively. Interestingly, hU-II(3-11) was 10 times more potent than the whole molecule (0.28 versus 3.3 nM for hU-II) but displayed only partial agonism. In addition, the removal of valine-11 (compounds 2 and 7) led to a dramatic loss of activity compared with the parent molecule. The Ala-scan series related to the exocyclic amino acid residues of hU-II showed that most derivatives are potent analogs of hU-II. Only the substitution of valine-11 (compound 17) gave rise to a 50-fold decrease in potency, accompanied with a reduced efficacy. On the other hand, an analog with a tetra-alanine N-terminal segment (compound 31: [Ala1,2,3,4]hU-II) or with the C-terminal carboxylic function transformed into an amidated form (compound 27) were partial agonists with an efficacy about half of that of hU-II but with EC50 values close to that of the parent molecule (compound 31: 6.0 ± 0.6 nM and compound 27: 2.0 ± 0.2 nM versus compound 1: 3.3 ± 0.3 nM).
Endocyclic and ring transformations of hU-II were also studied using the rat thoracic aorta bioassay (Table 3 and 4). So far, except for [Ala6]hU-II (compound 13), which was a weak agonist, all alanine substitutions of the endocyclic amino acids led to inactive peptides. Similarly, the successive deletion or the inversion of chirality of the endocyclic residues also exhibited, in most cases, a detrimental effect on the biological activity. One exception although with [d-Trp7]hUII (compound 23), which showed a significant but weak effect (EC50) on the aorta. Substitutions with isofunctional residues were carried out at various positions. For instance, lysine-9 was replaced with arginine (compound 32), and this analog was only 20 times less potent than hU-II. However, when the lysine-8 side chain was shortened by a distance equivalent to a methylene group, as obtained with ornithine-8 (compound 33), the new analog exhibited a complete loss of constricting activity. The role of the Phe-6 and Tyr-9 side chains were also investigated (compounds 35 and 36) using another aromatic residue for the substitution. Potent analogs with only a 10-fold decrease in activity were produced with this alteration.
Ring modifications were introduced in a few analogs (Table 4). The results showed that an increase of the ring size (compound 37) caused a 100-fold decrease in activity while linearization and substitution of the sulfhydryl moiety with acetamidomethyl groups (compound 34) gave a weakly active analog.
Binding Assays on Human and Rat Receptors. [125ITyr9]hU-II was used as a probe for evaluating the binding characteristics of the peptides on the human and rat U-II receptors (Fig. 2 for representative competition curves; Tables 5, 6, and 7). Nonspecific binding was determined in the presence of 10 μM human U-II and corresponded to 31 ± 2% (n = 5, mean ± S.E.M.) of the total binding for membranes expressing the human urotensin II receptor and to 8 ± 1% (n = 5, mean ± S.E.M.) of the total binding for membranes expressing the rat urotensin II receptor. The binding of [125ITyr9]urotensin II was characterized by a KD value of 1.4 ± 0.2 nM and a Bmax of 1.0 ± 0.1 pmol/mg of protein for human U-II receptor-expressing membranes (n = 4, mean ± S.E.M.) and a KD value of 1.25 ± 0.06 nM and Bmax of 3.4 ± 0.2 pmol/mg of protein for rat U-II receptor-expressing membranes (n = 5, mean ± S.E.M.). As in the rat thoracic aorta bioassay, human and rat U-II exhibited very similar affinities in the binding experiments. The magnitude of binding for both isoforms was superior on the human than on the rat receptor. Successive N-terminal deletions (Table 5; compounds 3 to 6) had only minor effects on the binding properties on both receptor types. Surprisingly, an exception was observed with hU-II(5-11) (compound 6) on the rat receptor since no binding was measured, although this peptide appeared as a weak agonist on the rat thoracic aorta. The removal of valine-11 in the whole molecule (compound 7) or in the hU-II(5-11) fragment (compound 2) led to molecules with poor affinities for the human and rat receptors. On the other hand, the substitution of residue 11 with alanine (compound 17) exhibited a 30-fold decrease on the rat receptor affinity and a 10-fold reduction of binding on the human receptor. All peptides with any other Ala substitution in the exocyclic segment of U-II (compounds 8 to 11), and even [Ala1,2,3,4]hU-II (compound 31), appeared as very potent ligands of the U-II receptors. Similarly, the C-terminal carboxylic function can be replaced with a neutral amide group without losing affinity for both human and rat receptors. Except for hU-II(5-11) (compound 6), which did not attach to the rat receptor but appeared as equipotent to hU-II in binding to the human receptor, all compounds containing exocyclic modifications exhibited a Hill slope close to -1, thus signifying the presence of a single U-II receptor population.
Endocyclic modifications confirmed the key role of the residues 6 to 9 of hU-II (Table 6). Successive deletions, inversion of amino acid configuration, or Ala substitutions gave rise to detrimental effects on the binding properties related to the human or rat U-II receptors. A few alterations were rather well tolerated as seen with the replacement of Tyr-9 with Phe (compound 35) or the substitution of Phe-6 with Tyr (compound 36). So far, both compounds exhibited very significant affinity for the U-II receptors.
Poor affinities were usually measured with peptide analogs containing cysteinyl modifications (Table 7). Nevertheless, the cyclic compound [HomoCys5-10]hU-II (compound 37) was described as a 100-fold weaker agonist than the parent molecule (EC50: 348 ± 23 versus 3.3 ± 0.3) in the rat thoracic bioassay. Strikingly, this analog showed no significant affinity for the rat receptor expressed in the HEK293 cell line but appeared well recognized by the human U-II receptor with an affinity about 100-fold lower than that of hU-II.
Functional FLIPR Assay on Ca2+Mobilization. Using the functional FLIPR-assay, the effects of exocyclic, endocyclic, and ring modifications on Ca2+ mobilization were measured on transfected CHO-K1 cells expressing the human or rat U-II receptors (Tables 8, 9, and 10). In contrast to the biological and binding results obtained with some exocyclic-altered analogs, the functional data showed that such modifications had in most cases only little effects on Ca2+ mobilization. In fact, EC50 values ranged from 1.8 to 193 nM with the human receptor, whereas it was between 0.24 to 735 nM with the rat receptor preparation. When endocyclic substitutions or mutations were introduced in hU-II, the absence of Ca2+ mobilization was, as expected, frequently encountered. Some transformations had little detrimental impacts on the calcium assay, however. This was particularly noticeable with the inversion of configuration of Trp-7 (compound 23; Table 9), the substitution of Tyr-9 with Phe (compound 35; Table 9), and the substitution of Phe-6 with Tyr (compound 36; Table 9). These analogs, as observed with the two other paradigms, retained on Ca2+ mobilization a very large activity. Surprisingly, analogs containing, respectively, a histidine residue at position 7 (compound 28; Table 9) or an ornithine moiety substituting Lys-8 (compound 33; Table 9) still exhibited a very significant potency in promoting calcium movements in the transfected cells. Similarly, the weak rat thoracic aorta agonist [HomoCys5-10]hU-II (compound 37) appeared as a very potent agonist on Ca2+ mobilization in CHO-K1 cells.
Discussion
Three series of urotensin II analogs were designed to elucidate the role of the exocyclic, endocyclic, and ring-related amino acids in three different paradigms, i.e., the rat thoracic aorta bioassay, binding assays with HEK293 cells expressing the human or rat U-II receptor, and functional assays on Ca2+ mobilization using CHO-K1 cells expressing the human or rat U-II receptor. The data revealed disparities depending of the assay or the species. For instance, the Ca2+ mobilization assay, whether carried out with the human or rat receptor (Table 8; compounds 2 to 6), suggested that the C-terminal part of U-II contains the pharmacophoric elements for giving agonists similar to the report of Itoh et al. (1987). However, the contractile effects evaluated in the rat aorta and the binding values measured with the human and rat receptors showed that the C-terminal fragments hU-II(5-11) (compound 6) and hU-II(5-10) (compound 2) were rather weak ligands or agonists. The structure of the cyclic core of urotensin II is similar to that of other peptides such as SST (Rohrer et al., 1998) and vasopressin (Manning et al., 1999). Therefore, it is not excluded that some U-II analogs and fragments might interact with receptors other than GPR14. As a matter of fact, although SST itself is not recognized by the U-II receptor, the group of Coy (Coy et al., 2000; Rossowski et al., 2002) screened a series of SST ligands in the rat aorta bioassay and for competitive binding in a rat preparation. They discovered that several analogs had moderate affinity for GPR14 and were able to block U-II induced phasic contractions in rat thoracic aorta rings. Coy et al. concluded that their data support the suggestion that U-II receptor and SST type 2/5 receptors display similar surface topologies.
Nevertheless, the rat aorta pharmacological assay and the binding studies showed correlations allowing the identification of useful pharmacophores. Hence, as suggested by Perkins et al. (1990), the “N-capping” of the C-terminal heptapeptide appears as a useful feature for obtaining potent peptides (compounds 3 to 5, 38, 8 to 11, 31, and rat U-II). Whether one amino acid or a short segment of residues is linked to the C-terminal heptapeptide, an effective analog can be obtained in such conditions. This strategy was used by a few groups (Kinney et al., 2001, 2002; Grieco et al., 2001), including ours (Flohr et al., 2002), by acetylating the N-terminal function of some U-II analogs. In contrast, as shown with [des-Val11]hU-II (compound 7) and [Ala11]hU-II (compound 17), the valine residue would play an important role in activity and affinity. It is worth mentioning that this amino acid is highly conserved throughout species, and the rare substitution observed for instance in the rat or carp is made with the isofunctional residue isoleucine (Conlon, 2000). Finally, it was demonstrated that the C-terminal negative charge of U-II is not important for the GPR14 receptor since its amidation (compound 27) did not significantly alter the affinity and biological activity of the U-II ligand.
Endocyclic modifications revealed the major role played by the Phe6-Trp7-Lys8-Tyr9 segment of hU-II. All assays (Tables 3, 6, and 9), including the Ca2+ mobilization paradigm, showed that deletions or substitutions in the cyclic core of the peptide gave rise frequently to a severe loss of activity and affinity, and as reported before (Flohr et al., 2002), this is particularly evident with the amino acid triad WKY. In fact, the substitution with Ala of the endocyclic residues led in most cases to the abolition of the contractile activity in the rat thoracic aorta. An exception was noticed with the evolutionarily highly conserved Phe-6, however, since [Ala6]hU-II retained some biological activity in the rat aorta. In agreement with these results, Phe-6 can be replaced with Tyr without losing much activity or affinity.
To explore the importance of the correct conformation of the message sequence, we inverted the chirality of the amino acids found in the cyclic part of hU-II (compounds 22 to 25). The incorporation of D-residues caused a major decrease of the agonistic activities and binding affinities, except for Trp-7. The corresponding analog [d-Trp7]hU-II exhibited significant activity and affinities, thus suggesting that the orientation is not critical as it was with the three other endocyclic residues. It is known that amino acids with a D-configuration support the formation of β-turn type structures, resulting in conformationally restricted peptides (Freidinger et al., 1984). Our previous NMR studies (Flohr et al., 2002) demonstrated, using the cyclic hexapeptide Ac-CF-wKYC-NH2, that the introduction of a d-Trp residue in position i + 1 gave a structure characterized by a distorted β-II′ turn. Therefore, we tried to favor the stabilization of an intramolecular β-turn conformation by introducing, according to the Chou-Fasman parameters (Chou and Fasman, 1978), a Pro moiety at position i + 1 (compound 41), or a sequential rearrangement of the endocyclic residues (compound 40). Although the Chou-Fasman <pt> parameter of these analogs was close to 1 and their respective Chou-Fasman pt value largely above the theoretical threshold limit of 0.75 × 10-4 (1.3 × 10-4 and 1.6 × 10-4 for [Lys7,Tyr8,Trp9]hUII and [Pro7]hU-II, respectively), thus exhibiting a strong probability of β-turn structure, both compounds were inactive in the biological and binding assays.
A few other alterations were also investigated. In occurrence, lysine-8 was replaced with the isofunctional amino acids Arg and Orn. It was observed that the rat thoracic aorta was very sensitive to those alterations since [Arg8]hU-II was 20-fold less potent than the parent molecule and [Orn8]hU-II was inactive. As concluded before (Flohr et al., 2002), it seems that the position and the distance of the positive ionizable feature, in relation to the two hydrophobic—aromatic poles formed by Trp-7 and Tyr-9 —are key characteristics of U-II ligands. It is likely, as described for Lys-9 of SST-14 (Marchese et al., 1995; Strnad and Hadcock, 1995), that Lys-8 of U-II would be involved in an interaction with GPR14, probably with the Asp-130 residue positioned in the TM3 of the receptor. Interestingly, when Lys-8 is substituted with Orn, the new derivative behaves as a competitive and selective antagonist in the rat aorta bioassay (Camarda et al., 2002).
In the third series of compounds, we explored the impact of modifications of the ring-related amino acids (Tables 4, 7, and 10). Hence, linearization of hU-II (analogs 12 and 34) caused a major loss of affinity and activity. Moreover, a peptide containing a Phe pair (compound 39), designed to mimic the π-π overlap of the aromatic rings observed in somatostatin, did not exhibit any significant binding nor activity. As demonstrated with [HomoCys5-10]hU-II, an analog with a disulfide-bridged ring containing two extra methylene moieties (compound 37), the ring size also appeared as a critical feature since this analog was about 100 times less active than U-II in the rat thoracic aorta assay. Similar conclusions were reached by Grieco et al. (2002) after exploring the ring size of hU-II analogs containing a lactam bridge. Interestingly, the affinity of compound 37 was significant in only one of the two binding paradigms. In this case, [HomoCys5-10]hU-II was only about 100-fold less potent than the parent molecule toward the human receptor, whereas it showed no affinity for rat GPR14.
In summary, this SAR study revealed the role of the amino acids of U-II in aorta contraction and human-rat receptor affinity and, more particularly, of the peptide triad WKY that appeared as the essential part of the physiological ligand. A few other amino acids might also be key residues for activity and/or affinity. The occasional lack of relationship between the results of the rat aorta paradigm and the Ca2+ mobilization functional assay sometimes led to difficult interpretations. This situation might be a consequence of the differences in Bmax values of the cell systems expressing human or rat receptors, as different expression levels may influence agonist potencies and efficacies. This phenomenon, described as the stimulus response coupling (Kenakin, 2002), was reported before (Kenakin and Beek, 1984; Rizzi et al., 1999) and could apply to our paradigms because the rat aorta expresses 250-1000 times less receptors than our transfected cell preparations (Bmax for the rat aorta according to Ames et al., 1999; 2-20 fmol/mg of protein, Bmax for our human U-II receptor-expressing membranes: 1 pmol/mg of protein, and Bmax for our rat U-II receptor-expressing membranes: 3.4 pmol/mg of protein). On the other hand, as suggested by the low Hill coefficients measured with some U-II analogs, the divergence between the bioactivity and the Ca2+ might also be related to effects induction through more than one receptor/binding site, not necessarily related to GPR14. Such an hypothesis has been also postulated previously by Coy et al. (2000) after observing a clear dichotomy between the phasic and tonic rat aorta contractions induced with U-II. Consequently, the analysis of functional data of U-II and analogs, using a Ca2+ mobilization assay developed with transfected mammalian cells expressing the human or rat GPR14 receptor, must be carried out with care as it might lead to ambiguous SAR conclusions.
Acknowledgments
We thank Myriam Létourneau for technical assistance.
Footnotes
-
↵1 Present address: 7TM Pharma, Hoersholm, Denmark and Laboratory for Molecular Pharmacology, The Panum Institute, University of Copenhagen, Denmark.
-
↵2 Present address: Novartis Pharma AG, Klybeckstrasse 141, Basel, Switzerland.
-
Financial supports were obtained from the Canadian Institutes for Health Research (CIHR) and the Heart and Stroke Foundation of Canada (HSFC). A.F. is “Chercheur National” from the Fonds de la recherche en santé du Québec (FRSQ). A.B. and S.B. are recipients of a studentship from the Natural Sciences and Engineering Research Council of Canada (NSERC).
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.103.052415.
-
ABBREVIATIONS: U-II, urotensin II; SST, somatostatin; SAR, structure-activity relationship; TFA, trifluoroacetic acid; FCS, fetal calf serum; PCR, polymerase chain reaction; DMEM, Dulbecco's modified Eagle's medium; HPLC, high-performance liquid chromatography; FLIPR, fluorescence imaging plate reader.
- Received April 1, 2003.
- Accepted May 28, 2003.
- The American Society for Pharmacology and Experimental Therapeutics