Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The 28-amino-acid neuropeptide vasoactive intestinal polypeptide (VIP) acts through two distinct receptors named VPAC₁ and VPAC₂ (Harmar et al., 1998). These two G protein-coupled, seven-TM-helix receptors have only 51% similarity (Couvineau et al., 1994; Svoboda et al., 1994) but recognize with a comparable, high-affinity VIP and the pituitary adenylate cyclase-activating polypeptide (PACAP). VPAC₁ and VPAC₂ receptors are differentially distributed in tissues (Ishihara et al., 1992; Usdin et al., 1994; Vertongen et al., 1997) and cell lines (Vertongen et al., 1996).

Two selective VPAC₂ receptors agonists were recently discovered: Ro 25-1553 (Gourlet et al., 1997a) and Ro 25-1392 (Xia et al., 1997) are cyclic analogs that differ from VIP by acetylation of the amino terminus, some mutations in the core of the peptide, the presence of a lactam bridge, and two additional lysine residues in the carboxyl terminus. There are no known selective high-affinity antagonist for the VPAC₂ receptors.

Two selective VPAC₁ receptor agonists have also been recently developed: one of these agonists is a derivative of the chicken secretin molecule, and the second, a hybrid peptide with the VIP amino terminus and a growth hormone-releasing peptide (GRF)-like carboxyl-terminal sequence (Gourlet et al., 1997b). A selective VPAC₁ receptor antagonist was also obtained through the single substitution of the VIP/GRF hybrid peptide (Gourlet et al., 1997c).

The group of Gozes recently described lipophilic VIP derivatives (Gozes and Fridkin, 1992; Gozes et al., 1994) and observed that the lipophilic VIP derivative obtained through the amidation of the amino terminus of VIP with an 18-carbon fatty acid (stearyl VIP) was 100-fold more potent than VIP in promoting neuronal cell survival in dissociated spinal cord cells (Gozes et al., 1995) but did not increase cAMP levels. We observed that this compound recognizes the recombinant rat and human VPAC₁ and VPAC₂ receptors with an affinity comparable to that of VIP and with a preference for the VPAC₂ over the VPAC₁ receptor but behaves like a partial agonist on adenylate cyclase stimulation (Gourlet et al., 1998). Related compounds like myristoyl- and palmitoyl-VIP also had a low intrinsic activity and behaved as VPAC₂-preferring partial agonists and/or competitive antagonists. These results suggested that acylation of the amino terminus might contribute to the VPAC₂ receptor selectivity of lipophilic VIP derivatives, Ro 25-1553 and Ro 25-1392, but that the length of the alkyl chain acylating the amino-terminal histidine residue may influence the intrinsic activity of the peptides.

These results prompted us to synthesize other derivatives, acylated at the amino terminus but with a shorter acyl chain. The hexanoyl-VIP (C₆-VIP) described in the present study had a 4-fold lower and 3-fold higher affinity than VIP for the human VPAC₁ and VPAC₂ receptors, respectively, and stimulated maximally the adenylate cyclase activity. The binding and functional properties of this new VPAC₂ preferring agonist were compared with the properties of Ro 25-1553 and acetyl-His¹- (AcHis¹)-VIP. By testing the ability of these agonists to recognize chimeric receptors, consisting of the interchange of the different domains of one receptor subtype grafted onto the core of the other receptor subtype, we were able to establish that the VPAC₂ selectivities of Ro 25-1553 and C₆-VIP were supported by different domains of the receptors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

All of the experiments were conducted on Chinese hamster ovary (CHO) cell membranes expressing the recombinant wild-type human VPAC₁, VPAC₂, or mutated receptors. The cell lines expressing the VPAC₁ and VPAC₂ receptors have been detailed previously (Ciccarelli et al., 1994; Svoboda et al., 1994). All chimeric receptors were constructed by the PCR overlap extension strategy. Briefly, cDNA fragments overlapping at their 5' or 3' extremity were generated using human VPAC₁ and VPAC₂ receptor cDNA as templates and appropriate chimeric primers. After purification of these fragments using the High Pure PCR Product Purification Kit (Boehringer-Mannheim Biochemica, Mannheim, Germany), they were used in a round of PCR overlap extension. The use of a phosphorylated forward primer surrounding the ATG initiation codon allowed us to obtain a 5' hemiphosphorylated cDNA fragment. This particularity combined with the presence of a 3'-A overhang resulting from the terminal transferase activity of Taq polymerase, allowing the unidirectional cloning of the chimeric receptors in pCR 3.1-Uni (InVitrogen, San Diego, CA), suitable for both prokaryotic and eukaryotic expressions. The first series of chimeric receptor was generated by the substitution of different domains in the VPAC₂ receptor by the counterpart of VPAC₁ receptor. N-VPAC₁/VPAC₂ chimera was generated by combining codons 1 to 143 of the VPAC₁ receptor with codons 128 to 438 of the VPAC₂ receptor. The NEC₁-VPAC₁/VPAC₂ combines codons 1 to 216 of the VPAC₁ receptor with codons 204 to 438 of VPAC₂ receptor. The receptor chimeric NEC₂-VPAC₁/VPAC₂ combines the codons 1 to 293 of VPAC₁ receptor with the codons 281 to 438 of VPAC₂ receptor. The latest chimeric receptor of this series (NEC₃-VPAC₁/VPAC₂) was generated by combining codons 1 to 373 of the VPAC₁ receptor with the codons 361 to 438 of the VPAC₂ receptor. The other series of chimeric receptor was generated by the substitution of the different domain in the VPAC₁ receptor by the counterpart of VPAC₂ receptor. The N-VPAC₂/VPAC₁ construction combined the VPAC₂ receptor codons 1 to 127 with the 144 to 457 VPAC₁ receptor codons. The NEC₁-VPAC₂/VPAC₁ construction was generated by combining the codons 1 to 203 of VPAC₂ receptor and the codons 217 to 457 of VPAC₁ receptor. The NEC₂-VPAC₂/VPAC₁ chimera combined the codons 1 to 280 of the VPAC₂ receptor with the codons 297 to 457 of VPAC₁ receptor. And the last chimeric receptor (NEC₃-VPAC₂/VPAC₁) has the codons 1 to 360 of the VPAC₂ receptor and the codons 374 to 457 of the VPAC₁ receptor. The polymerase chain reactions were performed using the Expand Long Template system (Boehringher, Mannheim) in the Geneamp 2450 thermocycler (Perkin-Elmer Cetus, Norwalk, CT). All sequences were verified using ABI Prism By Dye Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems, Norwalk, CT).

The methodology for cell transfection, as well as the cell culture medium, has been detailed previously (Van Rampelbergh et al., 1996). The cell clones expressing the different constructions were selected by testing the ability of the 10 µM VIP to stimulate the adenylate cyclase. One of the constructs (NEC₂-VPAC₁/VPAC₂) could not be identified by this method, so we also attempted to detect its expression in binding studies with ¹²⁵I-VIP, ¹²⁵I-C₆-VIP, and ¹²⁵I-VIP₁ antagonist with negative results.

Membranes were prepared from scraped cells lysed in 1 mM NaHCO₃ solution and immediate freezing in liquid nitrogen. After thawing, the lysate was first centrifuged at 4°C for 10 min at 400g; the supernatant was further centrifuged at 20,000g for 10 min. The pellet, resuspended in 1 mM NaHCO₃, was used immediately as a crude membrane fraction.

Binding studies were performed as described previously (Gourlet et al., 1997a) using either ¹²⁵I-VIP, ¹²⁵I-Ro 25-1553, or ¹²⁵I-C₆-VIP. The three tracers were radiolabeled similarly and had comparable specific radioactivity (Gourlet et al., 1997b). In all cases, the nonspecific binding was defined as residual binding in the presence of 1 µM VIP. Binding was performed at 37°C in a total volume of 120 µl containing 20 mM Tris-maleate, 2 mM MgCl₂, 0.1 mg/ml bacitracin, and 1% BSA (pH 7.4) buffer. From 3 to 30 µg of protein was used per assay. Bound and free radioactivity were separated by filtration through glass-fiber GF/C filters presoaked for 24 h in 0.01% polyethyleneimine and rinsed three times with a 20 mM (pH 7.4) sodium phosphate buffer containing 1% BSA.

Adenylate cyclase activity was determined according to the procedure of Salomon et al. (1974). Membrane proteins (3-15 µg) were incubated in a total volume of 60 µl containing 0.5 mM [-³²P]ATP, 10 µM GTP, 5 mM MgCl₂, 0.5 mM EGTA, 1 mM cAMP, 1 mM theophylline, 10 mM phospho(enol)pyruvate, 30 µg/ml pyruvate kinase, and 30 mM Tris · HCl at final pH 7.8.

The peptides used were synthesized in our laboratory as described previously (Gourlet et al., 1997b, 1998). The 1-hydroxybenzotriazole derivative of hexanoic acid was coupled to the amino terminus of VIP before cleavage and deprotection. The peptide purity was assessed by capillary electrophoresis, and the conformity was assessed by electrospray mass spectrometry.

All competition curves and dose-effect curves were analyzed by a nonlinear regression program (Prism; GraphPAD, San Diego, CA). The differences between the IC₅₀, EC₅₀, and maximal efficacy values were tested for statistical significance by Student's t test; P < .05 was accepted as being statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Properties of C₆-VIP on Recombinant VPAC₁ and VPAC₂ Receptors. On VPAC₂ receptors, the inhibition curves were identical when ¹²⁵I-VIP and ¹²⁵I-Ro 25-1553 were used as tracers (data not shown). AcHis¹-VIP and C₆-VIP were 6- to 4-fold less potent than VIP in binding studies on human recombinant VPAC₁ receptors, but 2- to 3-fold more potent than VIP in binding studies on recombinant VPAC₂ receptors, respectively (Table 1). Ro 25-1553 was 1000-fold less potent and 4-fold more potent than VIP on the VPAC₁ and VPAC₂ receptors, respectively.

View this table: [in this window] [in a new window]   TABLE 1 IC50 and EC50 values of tracer binding inhibition and adenylate cyclase activation on the recombinant human VPAC1, VPAC2, N-VPAC2/VPAC1, N-VPAC1/VPAC2, and N  EC1 VPAC1/VPAC2 receptors The values (±S.D.) were the mean of at least three determination.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Inhibition of tracer binding (top) and adenylate cyclase activation (bottoms) by VIP (), C6-VIP (), and Ro 25-1553 () using CHO cell membranes expressing the recombinant VPAC1 (A and D), N-VPAC2/VPAC1 chimera (B and E), and NEC1-VPAC2/VPAC1 chimera (C and F). 125I-VIP and 125I-Ro 25-1553 were used to characterize the VPAC1 receptors and the two chimeric receptors, respectively. The curves are the mean of at least three experiments performed in duplicate.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Inhibition of tracer binding (top) and adenylate cyclase activation (bottom) by VIP (), C6-VIP (), and Ro 25-1553 () using CHO cell membranes expressing the NEC2-VPAC2/VPAC1 chimera (A and D), NEC3-VPAC2/VPAC1 chimera (B and E), and the recombinant VPAC2 receptor (C and F). 125I-Ro 25-1553 were used to characterize these receptors. The curves are the mean of at least three experiments performed in duplicate.

Contribution of Amino-Terminal Portion and TM5 to EC₃ Region of Receptor to VPAC₂ Receptor Preference of Ro 25-1553 and C₆-VIP, Respectively. In a first set of experiments, we evaluated the properties of a chimeric receptor consisting in the amino-terminal (extracellular) VPAC₂ receptor domain followed by the seven-TM-helix domain of the VPAC₁ receptor (N-VPAC₂/VPAC₁). On that chimeric receptor, IC₅₀ values of binding inhibition were 0.5, 3.0, 5.0, and 5.0 nM for VIP, AcHis¹-VIP, C₆-VIP, and Ro 25-1553, respectively (Fig. 1B and Table 1). Thus, the IC₅₀ values of VIP, AcHis¹-VIP, and C₆-VIP were comparable to their VPAC₁ receptor IC₅₀ values, and that of Ro 25-1553 was comparable to its VPAC₂ receptor IC₅₀ value. In addition, the maximal stimulatory effect of Ro 25-1553 was higher than those of VIP, AcHis¹-VIP, and C₆-VIP. Thus, the amino-terminal domain of the VPAC₂ receptor was essential for high-affinity Ro 25-1553 recognition but not for that of C₆-VIP. In an attempt to identify the receptor region that supported the preferential C₆-VIP > VIP recognition by the VPAC₂ receptor, we tested additional chimeric receptors increasing the contribution of the VPAC₂ receptor sequence (Figs. 1C and 2, A and B). As shown in Figs. 1 and 2, chimeric receptors with the amino-terminal to EC₂-VPAC₂ receptor sequence (Fig. 2A) had, like VPAC₁ receptor, a preference for VIP over C₆-VIP. In contrast, the chimeric receptor with the amino-terminal to EC₃-VPAC₂ receptor sequence (Fig. 2B) preferred (like the VPAC₂ receptor; Fig. 2C) C₆-VIP over VIP. These results suggested that the hexanoyl anchoring point is somewhere between TM5 and EC₃.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Inhibition of tracer binding (top) and adenylate cyclase activation (bottom) by VIP (), C6-VIP (), and Ro 25-1553 () using CHO cell membranes expressing the recombinant chimeric receptors N-VPAC1/VPAC2 (A and D), NEC1-VPAC1/VPAC2 chimera (B and E), and NEC3-VPAC1/VPAC2 chimera (C and F) receptors. 125I-C6-VIP was used to characterize these chimeric receptors. The curves are the mean of at least three experiments performed in duplicate.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The natural ligands VIP and PACAP(1-27) have a clear preference for VPAC₁ over VPAC₂ receptors. In contrast, Ro 25-1553 and Ro 25-1392 are highly selective for VPAC₂ receptors. They differ from VIP by the acetylation of the amino-terminal histidine, several mutations in the central part of the molecule, a lactam ring between residues 21 and 25, and a basic carboxyl-terminal tail. Each modification could a priori contribute to the high affinity for the VPAC₂ receptor and the low affinity for the VPAC₁ receptor.

The following experimental data support the hypothesis that acylation of the amino terminus of these peptides probably does contribute to their VPAC₂ selectivity. 1) AcHis¹-VIP was reported to have a 2-fold higher affinity than VIP on the "helodermin preferring receptor" from the SUP-T1 lymphoblastic cell line (Svoboda et al., 1994), a receptor subsequently identified as the human VPAC₂ receptor (Robberecht et al., 1996), and a 5-to 10-fold lower affinity than VIP on the cloned rat VPAC₁ receptor (Gourlet et al., 1996a) (these data were confirmed in the present report on recombinant human VPAC₁ and VPAC₂ receptors: Table 1). 2) Three fatty acyl derivatives of VIP (myristoyl-, palmitoyl-, and stearyl-[N,Leu¹⁷]VIP), obtained through the conjugation of a fatty acid to the amino terminus of the peptide chain by an amide bond, also had a lower and higher affinity than VIP for the human VPAC₁ and VPAC₂ receptors, respectively. These three modified peptides were, however, partial agonists on the VPAC₁ and VPAC₂ receptors. Their intrinsic activity was in fact extremely low on the VPAC₂ receptors, where they behaved like partial agonists (Gourlet et al., 1998). These results prompted us to test the properties of an amino-terminally acylated VIP derivative modified with hexanoic acid. The C₆-VIP derivative had, compared with VIP, a 3-fold higher and a 4-fold lower IC₅₀ value in binding studies on VPAC₁ and VPAC₂ receptors, respectively. It behaved as a full and selective VPAC₂ agonist on adenylate cyclase stimulation.

We then attempted to identify the domains of the receptor implicated in the high- and low-affinity interaction of Ro 25-1553 and C₆-VIP for the VPAC₂ and VPAC₁ receptors, respectively. The results obtained with the first N-VPAC₂/VPAC₁ chimera, in the amino-terminal part of the VPAC₂ receptor grafted onto the VPAC₁ receptor TM domain, indicated clearly that the VPAC receptor amino-terminal domain is responsible for selective Ro 25-1553 recognition. This is not the case for AcHis¹-VIP and C₆-VIP, which had an identical potency (lower than that of VIP) on both VPAC₁ and N-VPAC₂/VPAC₁ receptors. These results supported the hypothesis that the amino-terminal sequence of the ligands interacted with the core of the receptor, whereas the remainder of the molecule, probably starting around amino acid 9 (the first mutated residue in Ro 25-1553) interacts with the extracellular amino-terminal domain of the receptor. These conclusions are in line with several observations on closely related receptors. 1) The amino-terminal domain of the secretin receptor is essential for the high-affinity interaction of secretin with its receptor (Vilardaga et al., 1995), and recognizes amino acids 8 to 15 (Gourlet et al., 1996b) and 22 (Dong et al., 1999) of secretin; the amino-terminal portions of the PACAP receptor (Van Rampelbergh et al., 1996; Hashimoto et al., 1997) and of the parathyroid hormone/parathyroid hormone-related peptide receptor (Zhou et al., 1997) also participate in the selective recognition of their respective ligands. 2) The amino terminus of secretin, and particularly the aspartate residue in position 3, interacts with several receptor residues located in the first and second TM domains (Vilardaga et al., 1996; Di Paolo et al., 1998, 1999).

We should very much like to identify the receptor region that recognizes the amino-terminal VIP histidine because this interaction is very important for receptor activation. Unfortunately, this histidine residue is likely also involved in stabilizing the active VIP conformation: conceiving VIP analogs that retain the right conformation but interact differentially with the receptor histidine anchoring site is by no means a trivial problem. We were therefore extremely interested by the observation that C₆-VIP is a full agonist, with a greater affinity than VIP for VPAC₂ receptor and only slightly lower affinity than VIP for VPAC₁ receptor. The VPAC receptor region that is responsible for the differential VIP/C₆-VIP recognition must be very close to the histidine anchoring site.

VIP was more potent than C₆-VIP on the NEC₂-VPAC₂/VPAC₁ chimeric receptor but less potent than VIP on the NEC₃-VPAC₂/VPAC₁ chimeric receptor and on the VPAC₂ receptor: the hexanoyl group probably recognized a binding site situated somewhere between TM5 and EC₃, and the amino-terminal histidine binding site (which is essential for VPAC₁ receptor activation) should be sought in the same region or in its immediate vicinity.

To further support this hypothesis, we attempted to express and study the binding and functional properties of the "mirror image" chimeras. Our results confirmed that the VPAC₂ TM7 sequence is not sufficient to support preferential C₆-VIP > VIP recognition (Fig. 3, C and F). We were, unfortunately, unable to detect the expression of one of these chimeras cloned in CHO cells (NEC₂-VPAC₁/VPAC₂) by functional or binding studies, and VIP had a surprisingly low affinity and efficacy on the two remaining chimeras (Fig. 2, A, B, D, and E). We subsequently observed that C₆-VIP had a higher affinity and efficacy than VIP on these chimeras (Table 1). Taken together, these results suggested that they could not be fully activated by the natural agonist, VIP, and that acylation with a 6-carbon acyl chain was sufficient to rescue the ability of the peptides to stabilize the active receptor conformation.

As pointed out recently by Colquhoun (1999), it is not possible to evaluate separately the affinity of agonists for the resting inactive receptors and their effect on receptor activation. If we assume that the agonist initially recognizes inactive receptors and then induces a conformational change to the active receptor conformation:
where K₁ and K₂ represent the affinity constant for the agonists for R and the agonist-receptor complex activation constant, respectively, then the affinity of the agonists measured in binding and functional studies is K_a = K₁(1 + K₂). In contrast, because antagonists recognize the inactive receptor conformation but do not induce receptor activation:
their affinity (K_a = K₁) does reflect their ability to recognize the receptor.

Removal of the amino-terminal VIP amino acids, acylation of the His¹ by long-chain fatty acids, or replacement of amino acids 2, 4, or 6 by D-amino acids markedly reduced not only the affinities of the peptides but also their intrinsic efficacies at VPAC₁ and VPAC₂ receptors: in the case of VIP related peptides, K₁ probably reflects the peptide anchoring through the central and carboxyl-terminal amino acids, and K₂, additional interactions between the amino-terminal VIP amino acids and the receptor in its active conformation.

To test whether the two VPAC₁/VPAC₂ chimeric receptors had a normal conformation in the resting state, we measured their affinities (K₁) for two antagonists: the VPAC₁ antagonist ([AcHis¹,D-Phe²,Lys¹⁵,Arg¹⁶,Leu²⁷]VIP(3-7)/GRF(8-27)] and VIP(5-28). The selectivity of the VPAC₁ antagonist is conferred mainly by its the carboxyl-terminal (8-27) region, that is, a peptide region thought to interact with the extracellular amino-terminal receptor domain. Our results, indicating that it had a high affinity for the N-VPAC₁/VPAC₂ chimeric receptor and low affinity for the N-VPAC₂/VPAC₁ chimeric receptor, further supported this hypothesis. They also suggested that the resting conformations of the amino-terminal domains of the VPAC₁/VPAC₂ chimeric receptors were normal or almost normal, because the antagonist affinities (K₁) were only slightly lower than those of the VPAC₁ receptor. In contrast, the ability of the TM chimeric receptors domain to recognize the amino-terminal VIP sequence and to change conformation in response to agonist recognition (K₂) was reduced, so we cannot exclude the hypothesis that the chimeric receptor TM domain had an altered conformation or was misoriented with respect to the amino-terminal domain. Merely acylating the amino-terminal histidine residue with an hexanoyl group (equivalent in size to a Leu or an Ile) was sufficient to "rescue" the affinity of the agonists and their ability to activate the receptor (K₂), suggesting that the energy input necessary to activate the receptor increased only slightly in the chimeric receptors compared with VPAC₁ and VPAC₂ receptors.

To conclude, our results suggested that 1) that the selectivity of the selective VPAC₂ agonist, Ro 25-1553, and of the selective VPAC₁ antagonist was supported by the extracellular amino-terminal receptor domain. The ability of the receptor to recognize preferentially either VIP or the C₆-VIP analog, in contrast, depended on the sequence of the carboxyl-terminal TM5 to EC₃ receptor domain. 2) The low affinity of the chimeric N-VPAC₁/VPAC₂ and NEC₁-VPAC₂/VPAC₁ receptors for VIP and AcHis¹-VIP reflected the reduced ability of these peptides to activate the chimeras rather than their inability to recognize the resting receptor conformation. 3) Acylation of the amino-terminal VIP histidine by a hexanoyl chain restored the ability of the peptides to activate these chimeric receptors. 4) Radioiodinated C₆-VIP was a good ligand for VPAC₁, VPAC₂, and chimeric receptor identification.