Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Corticotropin-releasing factor (CRF), a 41-amino acid peptide originally isolated from hypothalamus (Vale et al., 1981), is the main integrator of the stress response (Dunn and Berridge, 1990; Arborelius et al., 1999; Hauger and Dautzenberg, 1999). Central and peripheral effects of CRF and its structurally related analogs urocortin (UCN) (Vaughan et al., 1995; Donaldson et al., 1996), fish urotensin I (Lederis et al., 1982), and frog sauvagine (Montecucchi and Henschen, 1981) are mediated by their binding and activation of two CRF receptors (CRF₁ and CRF₂), which belong to the class B subfamily of G protein-coupled receptors (Vale et al., 1997). CRF₁ and CRF₂ receptors are ~70% homologous and couple to stimulatory GTP-binding proteins (reviewed in Dautzenberg et al., 2001a). Three biologically active splice variants (CRF_2A-C) have been identified for the CRF₂ receptor (see Kilpatrick et al., 1999).

Despite a high degree of sequence homology, the specificity of CRF agonist binding to CRF₁ and CRF₂ proteins differs to a considerable extent. The mammalian CRF₁ receptor nonselectively recognizes CRF, UCN, urotensin I, and sauvagine. These four CRF peptides bind to the CRF₁ receptor with similar degrees of high affinity and equipotently stimulate intracellular cAMP accumulation (Vaughan et al., 1995; Donaldson et al., 1996; Dautzenberg et al., 1997, 1999; Palchaudhuri et al., 1998). In contrast, the Xenopus laevis CRF₁ receptor (xCRF₁) selectively binds CRF agonists in a highly selective manner whereby human CRF (hCRF), X. laevis CRF (xCRF), urotensin I, and rat UCN are recognized with a significantly higher affinity than the structurally related analogs ovine CRF (oCRF) and sauvagine (Dautzenberg et al., 1997).

The mammalian and X. laevis CRF₂ receptors display substrate specificities that differ from the mammalian CRF₁ and the xCRF₁ receptor (Donaldson et al., 1996; Dautzenberg et al., 1997, 1999; Ardati et al., 1999; Palchaudhuri et al., 1999). The CRF peptides hCRF, oCRF, and xCRF bind with significantly lower affinity than UCN, urotensin I, and sauvagine.

The identification of regions forming the binding pocket or being critical for ligand selectivity of the mammalian CRF₁, xCRF₁, and human CRF_2A receptors has been the subject of various studies (Liaw et al., 1997a,b; Dautzenberg et al., 1998, 1999; Perrin et al., 1998; Wille et al., 1999; Assil et al., 2001). From those studies, it became evident that both receptors use amino acids that are within the extracellular (EC) domains of the receptor or at the interface between the EC domains and the transmembrane helices. The ligand-selective regions of the CRF₁ and CRF₂ receptor are located, however, in different regions of these two proteins. The major determinant for high-affinity ligand binding of the CRF₁ receptor resides in its N-terminal EC1 domain (Perrin et al., 1998; Wille et al., 1999; Assil et al., 2001), whereas the ligand-selective domains of human CRF_2A (hCRF_2A) have been identified in EC2 and at the junction of EC3 and transmembrane 5 (Liaw et al., 1997a,b). Replacement of the amino acids of hCRF₁ with residues at equivalent positions of hCRF_2A created a mutated receptor that no longer bound hCRF with high affinity (Liaw et al., 1997a). In agreement with these findings, the agonist-selective domains of the xCRF₂ receptor have been mapped to regions other than EC1, indicating that this receptor uses domains similar to its human CRF_2A counterpart (Dautzenberg et al., 1999). Recently, we reported the first evidence that the binding modes of mammalian CRF₂ receptors and the xCRF₂ receptor differed; furthermore, depending on the radioligand used, rank order binding profiles differed (Dautzenberg et al., 2001b). When competition binding experiments were performed with astressin, a nonselective antagonist (Gulyas et al., 1995), this ligand bound to the xCRF₂ receptor with an affinity more than 10-fold higher compared with all other CRF radioligands (Dautzenberg et al., 2001b). Because the ligand-selective domains mapped to the hCRF_2A receptor are not well conserved in the xCRF₂ receptor (Fig. 1), we speculated that one or more of these regions might be responsible for the differential binding profile of the xCRF₂ receptor.

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Two-dimensional model (A) and sequence comparison of the EC3 (B) domain of the hCRF2A and xCRF2 receptors. The conserved amino acids are presented as filled circles, whereas open circles represent divergent residues. The locations of the ligand-selective amino acids of the hCRF2A receptor are highlighted. Sequences surrounding the ligand-selective amino acids in EC3 (B) are shown, and the important triplet of both receptors is highlighted.

In this study, we generated mutated hCRF₂ and xCRF₂ receptors and tested their binding profile using the radioligands ¹²⁵I-sauvagine and ¹²⁵I-astressin. In addition, we compared the ability of astressin and anti-sauvagine-30 (aSVG) to inhibit agonist-mediated cAMP accumulation in HEK293 cells stably transfected with the human and amphibian wild-type or mutant receptors.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials, Peptides, and Reagents. All cell culture media and reagents were purchased from Invitrogen (Basel, Switzerland). The CRF peptides (purity, >95%) were obtained from Bachem Corp. (Bubendorf, Switzerland), whereas aSVG (purity, >95%) was synthesized in-house according to a previously published method (Rühmann et al., 1996; Brauns et al., 2001).

Radiochemicals. ¹²⁵I-Astressin (2200 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA), whereas ¹²⁵I-Tyr⁰-sauvagine (¹²⁵I-sauvagine; 2000 Ci/mmol) was purchased from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK).

Construction of Mutated Receptors. The residues Glu⁶⁶Tyr⁶⁷ and Asp²⁶²Leu²⁶³Val²⁶⁴ of the hCRF_2A receptor were mutated into the corresponding amino acids of the xCRF₂ receptors (Asp⁶⁹Ser⁷⁰ and Lys²⁶²Tyr²⁶³Ile²⁶⁴) to create the receptor mutants hR2DS, hR2KYI, and hR2DS-KYI. The residues Asp⁶⁹Ser⁷⁰ and Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ of the xCRF₂ receptor were mutated to the amino acids occupying these positions in the hCRF_2A receptor to generate the xR2EY, xR2DLV, and xR2EY-DLV mutants. Mutagenesis was accomplished using the QuickChange kit (Stratagene, La Jolla, CA) as reported previously (Dautzenberg et al., 1998; Wille et al., 1999). The wild-type and mutant receptors were cloned into the pcDNA3 vector (Invitrogen). Sequences obtained by polymerase chain reaction were verified by DNA sequencing using an ABI 310 DNA sequencer (Applied Biosystems, Weiterstadt, Germany); the GCG software package (Accelrys, Cambridge, UK) was used for analysis.

Cell Transfections and Radioreceptor Binding Assays. cDNAs of hCRF_2A, xCRF₂, and mutated receptors, all inserted into the pcDNA3 vector (2 µg each), were stably transfected into HEK293 cells with the Transfectam reagent (BioSepra, Inc., Villeneuve La Garenne, France) as reported previously (Dautzenberg et al., 2000). Two days after transfection, geneticin selection (500 µg/ml) was initiated, and clones expressing low to moderate receptor levels (500-1200 fmol/mg) were selected.

cAMP Assays. HEK 293 cells, stably expressing hCRF_2A, xCRF₂, or mutated receptors, were plated at 10,000 to 50,000 cells per well in 96-well dishes. Transfected cells were exposed to CRF peptides for a 10-min stimulation period at 37°C (5% CO₂). During the 10-min incubation period, no desensitization of the cAMP signal was observed. However, longer CRF stimulation periods resulted in considerable desensitization of the cAMP signal, especially at agonist concentrations higher than 100 nM (sauvagine and UCN; unpublished observations). The antagonistic experiments with aSVG and astressin were conducted by simultaneous application of the respective antagonist and sauvagine followed by a 10-min incubation as described above. Measurement of intracellular cAMP levels was performed as described previously (Dautzenberg et al., 2001c). The data were analyzed by two-way analysis of variance, and significance between groups was determined by post hoc analysis using Dunnett's test.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Binding Properties of hCRF_2A, xCRF₂, and Two Mutated CRF₂ Receptors. cDNAs were synthesized to encode the following two different mutations in the EC3 domain of the human and X. laevis CRF₂ receptors: hR2KYI (amino acids Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ of xCRF₂) and xR2DLV (residues Asp²⁶²Leu²⁶³Val²⁶⁴ of hCRF_2A). After these constructs were stably transfected into HEK293 cells, binding profiles using a radiolabeled agonist (¹²⁵I-sauvagine) and an antagonist (¹²⁵I-astressin) were determined. Because potential differences in the binding profiles of the mutated receptors could result from differences in G protein-coupling properties (see Kenakin, 1997), we excluded this possibility by assessing binding in the presence of increasing concentrations of GTPS or Gpp(NH)p. Both GTP analogs potently inhibited ¹²⁵I-sauvagine binding to the receptor preparations to 56 to 72% (Table 1). In addition, the GTP analog inhibited ¹²⁵I-sauvagine binding to the CRF₂ receptor preparation with similar affinity (Table 1). In agreement with its antagonist properties, binding of ¹²⁵I-astressin was not inhibited by Gpp(NH)p and GTPS (data not shown). Thus, we concluded that the native and mutant CRF₂ receptors exhibited the same degree of coupling to the endogenous Gs proteins.

View this table: [in this window] [in a new window]   TABLE 1 Inhibition of 125I-sauvagine binding to hCRF2A, xCRF2, and two mutated CRF2 receptors by GTPS and Gpp(NH)p The data are means ± S.E.M. from two different binding experiments performed in triplicate using 125I-sauvagine as radioligand.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Competitive binding of astressin to the hCRF2A, xCRF2, hR2KYI, and xR2DLV receptors in the presence of either 125I-sauvagine (A) or 125I-astressin (B) as radioligands. Membranes were incubated with increasing concentrations of the CRF antagonists for 120 min at 22°C. The results are representative of three independent binding experiments.

Binding Affinities of Human and Amphibian CRF₂ Receptors Carrying Point Mutations in the EC1 or EC1/EC3 Domains. Mutations in the EC3 domain did not completely reverse the differential binding profiles of the hCRF_2A and xCRF₂ receptor. Consequently, experiments were performed to compare binding affinities for ligands at human and amphibian CRF₂ receptors carrying point mutations restricted to the EC1 domain or a combination of EC1 and EC3 mutations. Because a two-amino acid motif in the ligand-selective domain of the xCRF₁ receptor differs in hCRF_2A (Glu⁶⁶Tyr⁶⁷) and xCRF₂ (Asp⁶⁹Ser⁷⁰) receptors (Dautzenberg et al., 1998), the following four mutated receptors were constructed: hR2DS (EC1 mutation), hR2DS-KYI (EC1/EC3 mutation), xR2EY (EC1 mutation), and xR2EY-DLV (EC1/EC3 mutation). After these mutated receptors were stably transfected into HEK293 cells, their binding properties were characterized using ¹²⁵I-sauvagine as the radioligand.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Competitive binding of astressin to the mutated CRF2 receptors hR2DS, hR2DS-KYI, xR2EY, and xR2EY-DLV using 125I-sauvagine as the radioligand. Membranes were incubated with increasing concentrations of the CRF antagonists for 120 min at 22°C. The results are representative of three independent binding experiments.

Stimulation of cAMP Accumulation in HEK293 Cells Expressing hCRF_2A, xCRF₂, and Mutated Receptors. The ability of CRF agonists to stimulate cAMP accumulation was assessed in HEK293 cells stably expressing native or mutated CRF₂ receptors. A similar potency rank order (sauvagine > UCN > hCRF > oCRF) was observed among the four receptors (Table 5). Because sauvagine stimulated intracellular cAMP accumulation with the greatest potency, we assessed the inhibitory effects of the nonselective CRF receptor antagonist astressin and the selective CRF₂ receptor antagonist aSVG on cAMP stimulation produced by 1 nM sauvagine, which is a concentration slightly above the EC₅₀ value observed in the four receptor lines (Table 5). When cells expressing the four different CRF₂ receptor proteins were coincubated with 100 nM astressin or aSVG, the sauvagine-stimulated cAMP accumulation was significantly reduced in the four receptor lines, indicating antagonist actions (Figs. 4 and 5).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Inhibition of sauvagine-mediated cAMP accumulation in HEK293 cells stably expressing the hCRF2A, xCRF2, hR2KYI, or xR2DLV receptors by aSVG astressin. Cells were incubated in the presence or absence of 1 nM sauvagine with or without 10 nM antagonist for 10 min at 37°C. Each bar represents the mean ± S.E.M. of three independent stimulation experiments conducted in quadruplicate.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Antagonist properties of astressin with the two natural CRF2 receptors hCRF2A and xCRF2 and two mutated receptors. Full dose-response curves for sauvagine-mediated cAMP accumulation were conducted in the presence of increasing concentrations (1-100 nM) of astressin. Data represent triplicates of one representative experiment repeated three times.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The purpose of this study was to identify the residues of the xCRF₂ receptor that govern its higher degree of binding selectivity compared with its human counterpart (Dautzenberg et al., 2001). Although the mammalian and amphibian CRF₁ receptor use amino acids located in the N-terminal EC1 domain for high-affinity ligand binding and substrate recognition (Dautzenberg et al., 1998; Perrin et al., 1998; Wille et al., 1999; Assil et al., 2001), the hCRF_2A and xCRF₂ receptor were reported to use different exofacial domains. For the hCRF_2A receptor, three regions within the EC2 and EC3 domain of this receptor have been identified to be critical for selective binding of and activation by CRF agonists (Liaw et al., 1997a). In our recent study, similar regions in EC2 and EC3 seemed to confer its agonist selectivity (Dautzenberg et al., 1999).

Interestingly, the amino acid motifs that probably mediate the ligand selectivity of hCRF_2A are not well conserved between the human and X. laevis receptors (Dautzenberg et al., 1997, 1999). Moreover, we have recently shown that a histidine residue reported to be located at position 185 and to play a crucial role for the binding specificity of hCRF_2A (Liaw et al., 1997a) most likely represents a polymerase chain reaction artifact. Instead, our sequencing experiments identified an arginine residue (Arg¹⁸⁵), located at position 185 of the hCRF_2A cDNAs, isolated from a variety of tissues, which is conserved in the hCRF_2A gene as well as other vertebrate CRF₁ and CRF₂ receptors (Dautzenberg et al., 2000). Furthermore, the second important domain for the ligand selectivity of the hCRF_2A receptor, residues Val¹⁷²Asp¹⁷³His¹⁷⁴ (Liaw et al., 1997a), are almost identical to the equivalent region of the xCRF₂ receptor (Ile¹⁷⁴Asp¹⁷⁵His¹⁷⁶) (Dautzenberg et al., 1997).

The above findings led us to first focus our investigation on the third ligand-selective domain, residues Asp²⁶²Leu²⁶³Val²⁶⁴ of the hCRF_2A receptor (Liaw et al., 1997a). The equivalent domain of the xCRF₂ receptor, residues Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ (Dautzenberg et al., 1997), differ strongly from its human counterpart.

Indeed, mutagenesis of these three amino acids altered the binding pharmacology of the two mutants. The xR2DLV mutant, encoding the residues of the hCRF_2A receptor, displayed the same binding preferences with both radioligands, ¹²⁵I-sauvagine and ¹²⁵I-astressin, as the human receptor. For the human mutant hR2KYI, the presence of the Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ triplet of the xCRF₂ receptor resulted in a pharmacology profile closely resembling the amphibian CRF₂ receptor. The largest differences in binding-affinity profiles were observed for the agonists hCRF and sauvagine and for the antagonists aSVG and astressin. Interestingly, the hR2KYI mutant and the xCRF₂ receptor bound hCRF with low affinity when ¹²⁵I-sauvagine was the competed radioligand. For the xR2DLV and hCRF_2A receptor, a low binding affinity for hCRF was observed in the presence of ¹²⁵I-astressin, consistent with the concept that binding of an agonist is reduced in the presence of a radiolabeled antagonist (Sleight et al., 1996; Dautzenberg et al., 2001b). Furthermore, sauvagine binding in the presence of ¹²⁵I-sauvagine occurred with a lower affinity to the xCRF₂ and hR2KYI receptor compared with other ligands, whereas sauvagine bound with high affinity to the hCRF_2A and xR2DLV receptor. However, in the presence of ¹²⁵I-astressin, the agonist showed no differences in binding to the four receptors. The CRF₂-selective antagonist aSVG, like sauvagine, bound with a lower affinity to the xCRF₂ and hR2KYI receptor than to the hCRF_2A receptor and the xR2DLV mutant. However, unlike classical antagonists, which bind independent of the agonistic or antagonistic nature of the competed radioligand (Sleight et al., 1996; Perrin et al., 1999), the affinity of aSVG was shifted to the right with all four receptors. This effect was strongest for the xCRF₂ receptor and hR2KYI mutant, which showed only a moderate affinity of ~20 nM for the antagonist. This unusual behavior for a receptor antagonist suggests that aSVG binding selectively depends on the agonistic or antagonistic nature of the competed radioligand. Nevertheless, both the nonradioactive and the iodinated form of the peptide have been shown to behave as antagonists (Rühmann et al., 1998; Higelin et al., 2001).

Notably, the mutagenesis approach within the EC3 domain did not unravel completely the unusual binding profile of astressin to the xCRF₂ receptor. Although incorporation of the amino acid triplet Asp-Leu-Val into the xCRF₂ receptor created a mutant with binding preferences indistinguishable from that of the hCRF_2A receptor, the converse result was not obtained when the corresponding human sequence was replaced by the triplet Lys-Tyr-Ile in the hR2KYI mutant. This mutant bound astressin with high affinity in the presence of ¹²⁵I-sauvagine and thus differed from the xCRF₂ receptor, which showed a lower affinity for astressin when ¹²⁵I-sauvagine was the competed radioligand (Dautzenberg et al., 2001b).

Thus, we concluded that an unidentified region of the xCRF₂ receptor interacts with the amino acid triplet Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ to negatively regulate astressin binding. Similarly, a recent study demonstrated that the binding of astressin differed from CRF and UCN at rat CRF₁ receptor (Perrin et al., 1998). Although astressin binding to the rat CRF₁ receptor only requires the EC1 domain, the EC4 domain is also needed for binding of CRF and UCN (Perrin et al., 1998). Furthermore, a microheterogeneity of agonist binding to the hCRF_2A receptor was also reported (Liaw et al., 1997b). Although introduction of the ligand-selective regions of the hCRF_2A receptor into the sequence of the hCRF₁ receptor strongly impaired CRF and sauvagine binding, UCN binding was insensitive to this mutation (Liaw et al., 1997b). These findings are not only restricted to the CRF receptor system because a similar observation was reported for another member of the class B G protein-coupled receptor subfamily, the parathyroid hormone receptor, which shows involvement of different amino acids for the binding of different parathyroid hormone variants (Lee et al., 1995).

We focused our search for the additional extracellular region that impairs binding of astressin to the xCRF₂ receptor on the N-terminal EC1 domain. Recently, we identified an amino acid doublet within the ligand-binding site of the mammalian CRF₁ receptor (Wille et al., 1999; Assil et al., 2001) in a parallel study investigating the site for astressin binding to the xCRF₁ receptor (S. Wille, J. Higelin, and F. M. Dautzenberg, manuscript in preparation). This doublet is conserved in xCRF₁ (Glu⁷⁰Tyr⁷¹) and hCRF_2A (Glu⁶⁶Tyr⁶⁷) but diverges in hCRF₁ (Ala⁷⁰Phe⁷¹) and the xCRF₂ receptor (Asp⁶⁹Ser⁷⁰). Interestingly, mutagenesis of this amino acid doublet did not alter binding of astressin to the hR2DS and xR2EY receptors. In contrast, replacement of EC1/EC3 residues of the hCRF_2A receptor with the corresponding X. laevis motifs resulted in a xCRF₂ pharmacology. Likewise, replacement of EC1/EC3 residues of the xCRF₂ receptor with the corresponding human motifs shifted its ligand rank order to an hCRF2A profile.

Rank orders for sauvagine-stimulated cAMP accumulation resembled rank orders obtained with binding studies where ¹²⁵I-sauvagine was used as the radioligand. Although astressin and aSVG were highly potent and competitive antagonists at the hCRF_2A, hR2DS, xR2DLV, and xR2EY-DLV receptors, both antagonists were significantly less potent at the hR2DS-KYI, xCRF₂, and xR2EY receptors. However, astressin was a more potent antagonist than aSVG in hR2KY-expressing cells in agreement with astressin possessing a higher binding affinity than aSVG at the hCRF_2A receptor. Thus, the binding properties of both antagonists correspond closely when ¹²⁵I-sauvagine rather than ¹²⁵I-astressin is used as the radioligand. Furthermore, these data demonstrate that the binding pocket of xCRF₂ for astressin differs from its binding pockets for agonists and other antagonists. Importantly, the slopes for the antagonism of aSVG and astressin of sauvagine-stimulated cAMP accumulation were ~1. Although our aSVG data agree with the findings of a recent study (Brauns et al., 2001), the slopes for astressin antagonism differ. Because our transfected cell lines express CRF₂ receptors at a level significantly lower than the CRF₂ receptor expression level used by Brauns et al. (2001), our data may reflect more native cell lines endogenously expressing CRF receptors.

In conclusion, our site-directed mutagenesis experiments have identified two important regions mediating the differential binding of CRF analogs to the amphibian CRF₂ receptor: a) an amino acid triplet Lys²⁶⁴Tyr²⁶⁵Ile²⁶⁶ in the EC3 domain and b) a two-amino acid motif, Asp⁶⁹Ser⁷⁰, in the EC1 domain of the xCRF₂ receptor. Replacement of this region by the corresponding amino acids of the hCRF_2A receptor (Glu⁶⁶Tyr⁶⁷ and Asp²⁶²Leu²⁶³Val²⁶⁴) generated a mutant with a binding pharmacology indistinguishable from that of the hCRF_2A receptor. The converse replacement of this region in the human CRF_2A receptor with the corresponding X. laevis sequence shifted the binding profile to that of the xCRF₂ receptor. Finally, antagonism of sauvagine-stimulated cAMP accumulation by astressin and aSVG followed competitive binding data using ¹²⁵I-sauvagine as a radioligand. Therefore, microheterogeneity within the ligand-binding pocket seems to be present in the amphibian CRF₂ receptor.