Co-evolution of ligand-receptor pairs in the vasopressin/oxytocin superfamily of bioactive peptides.

In order to understand the molecular mechanisms that underlie the co-evolution of related yet functionally distinct peptide-receptor pairs, we study receptors for the vasopressin-related peptide Lys-conopressin in the mollusc Lymnaea stagnalis . In addition to a previously cloned Lys-conopressin receptor (LSCPR1), we have now identified a novel Lys-conopressin receptor sub- type, named LSCPR2. The two receptors have a differential distribution in the reproductive organs and the brain, which suggests that they are involved in the control of distinct aspects of reproduction and mediate transmitter-like and/or modulatory effects of Lys-conopressin on different types of central neurons. In contrast to LSCPR1, LSCPR2 is maximally activated by both Lys-conopressin and Ile-conopressin, an oxytocin- like synthetic analog of Lys-conopressin. Together with a study of the phylogenetic relationships of Lys- conopressin receptors and their vertebrate counterparts, these synthetic analog Ile-conopres-sin. We show that vasopressin and oxytocin receptors evolved from an LSCPR2-like ancestral receptor, and propose a mechanism for the molecular evolution of specificity in the peptide-receptor pairs of the vasopressin/oxytocin superfamily. conserved of sequences. Bam several brain-specific in order obtain insert-containing vivo region and sequences endonucleases positive control for each cDNA preparation. After 30 cycles of amplification, PCR products were separated on agarose gel, transferred to Hybond N (cid:49) membrane (Amersham Corp.), hybridized to 5 (cid:57) -end 32 P-labeled primers specific for each product, and autoradiographed. In Situ Hybridization— Digoxigenin-labeled run-off antisense RNA was synthesized from 5 (cid:57) -end-linearized plasmids pBSCPR1 (26) and pBSCPR2 using 20 units of T3 RNA polymerase and digoxigenin-UTP labeling mixture (both from Boehringer Mannheim). Hybridizations of alternately collected sections of Lymnaea brains were performed as described previously (26). Phylogenetic Analysis— Sequences of the members of the vasopres-sin/oxytocin receptor family were extracted from a G protein-coupled receptor data base and aligned using the program Clustal V Alignments were adjusted manually, and the transmembrane domains, as defined by Baldwin (33), were utilized to generate a PAUP data matrix (34). An accepted mutation stepmatrix was calculated using the program AMP. A branch and bound search was performed without the stepmatrix, and the three most parsimonious trees with tree lengths of 858 were selected. When the branch and bound search was repeated with the stepmatrix invoked, one of these trees was found to be the shortest. Using the PAUP program, 1,000 replicates of bootstrap analysis were performed using the heuristic search procedure. The calcu- lated g 1 value of (cid:50) 0.869935 ( p (cid:46) netic

In order to understand the molecular mechanisms that underlie the co-evolution of related yet functionally distinct peptide-receptor pairs, we study receptors for the vasopressin-related peptide Lys-conopressin in the mollusc Lymnaea stagnalis. In addition to a previously cloned Lys-conopressin receptor (LSCPR1), we have now identified a novel Lys-conopressin receptor subtype, named LSCPR2. The two receptors have a differential distribution in the reproductive organs and the brain, which suggests that they are involved in the control of distinct aspects of reproduction and mediate transmitter-like and/or modulatory effects of Lysconopressin on different types of central neurons. In contrast to LSCPR1, LSCPR2 is maximally activated by both Lys-conopressin and Ile-conopressin, an oxytocinlike synthetic analog of Lys-conopressin. Together with a study of the phylogenetic relationships of Lysconopressin receptors and their vertebrate counterparts, these data suggest that LSCPR2 represents an ancestral receptor to the vasopressin/oxytocin receptor family in the vertebrates. Based on our findings, we provide a theory of the molecular co-evolution of the functionally distinct ligand-receptor pairs of the vasopressin/oxytocin superfamily of bioactive peptides.
Peptide receptors form an important subclass of the otherwise diverse superfamily of G protein-coupled 1 receptors that all have a seven-transmembrane segment topology in common (for review, see Ref. 1). This receptor type forms a component of a modular system for the transduction of extracellular signals across the cell membrane and the subsequent conversion of these signals to an intracellular second messenger pathway via the activation of heterotrimeric G proteins (for review, see Ref. 2). Molecular evolutionary mechanisms such as gene duplication and subsequent mutation of the resulting genes have resulted in the formation of families of related yet distinct peptides (3) and peptide receptors (4). In the nervous and endocrine systems, many peptide isoforms bind to distinct receptor subtypes that mediate the specific cellular actions that underlie a large variety of behavioral and physiological processes. Peptide-receptor and receptor-effector interactions are of critical importance for peptide function, and many diseases are linked to malfunctions of these interactions (5). However, the great variety in structure of both peptides and receptors has hampered the development of a coherent theory explaining the molecular basis of co-evolution of specifically interacting peptide-receptor pairs.
The vasopressin/oxytocin superfamily of peptides and their cognate receptors offer an attractive model for the study of specificity of peptide-receptor interactions and the co-evolution of related peptide-receptor pairs (6). These peptides and receptors occur throughout the entire animal kingdom (for review, see Ref. 7), allowing detailed comparison of their structural features and experimental testing of putative specificity determinants. The peptides of the superfamily are surprisingly alike, sharing at least 5 out of 9 residues and a disulfide-linked ring structure that put severe constraints on conformational flexibility (8,9). In the vertebrates, gene duplication gave rise to related yet distinct vasopressin and oxytocin genes (10). The different functions of vasopressin and oxytocin are mediated by genetically distinct receptor subtypes (11,12). The V2 vasopressin receptor mediates the antidiuretic effects of vasopressin and couples positively to adenylate cyclase (13)(14)(15)(16). By contrast, the V1a (17,18) and the V1b (19) vasopressin receptors, which mediate the effects of vasopressin on liver glycogenolysis and on adrenocorticotropin release, respectively, as well as the oxytocin receptor (15,20,21), which mediates the various central and peripheral functions in reproduction of oxytocin, all couple to the inositol trisphosphate/Ca 2ϩ signal transduction pathway. The discriminative binding of vasopressin and oxytocin to their respective receptors is dictated to a large extent by the amino acid residue at position 8. At this position, the family of vasopressin and related peptides have a basic residue, whereas oxytocin and related peptides have a neutral, in most cases aliphatic residue (7 In the invertebrates, on the other hand, only one gene is present that is considered the present-day representative of the ancestral peptide gene to the vasopressin/oxytocin superfamily (9). In the mollusc Lymnaea stagnalis, this gene encodes Lysconopressin, a peptide that is structurally related to vasopressin and that has a wide distribution throughout the invertebrate phyla (9,(22)(23)(24)(25). In Lys-conopressin, the chemical nature of the amino acid residue at position 8 is not important for peptide function, since replacement of this basic residue (lysine) by an aliphatic one (isoleucine) does not affect the potency of binding of the peptide to its receptor (9). These studies also show that Lys-conopressin, although structurally related to vasopressin, controls reproductive functions that are clearly analogous to the central and myoregulatory functions of oxytocin in vertebrate reproduction. In addition, Lys-conopressin serves vasopressin-like functions in the control of carbohydrate metabolism (26). Together, these observations suggest that functionally distinct vasopressin and oxytocin receptors may have evolved from a nondiscriminative Lys-conopressin receptor-like receptor by gene duplication and subsequent introduction of specificity determinants that enable discriminative binding of vasopressin and oxytocin on the basis of the chemical nature of amino acid residue 8.
We have recently cloned a Lys-conopressin receptor from the Lymnaea vas deferens, named L. stagnalis conopressin receptor (LSCPR), that displays discriminative binding properties with respect to residue 8 in Lys-conopressin (26). Here we report the cloning of a second Lys-conopressin receptor, named LSCPR2, that unlike the previously identified receptor (now renamed LSCPR1), does not discriminate between Lysconopressin and its oxytocin-like synthetic analog Ile-conopressin. We show that vasopressin and oxytocin receptors may have evolved from an LSCPR2-like ancestral receptor, and propose a mechanism for the molecular evolution of specificity in the peptide-receptor pairs of the vasopressin/oxytocin superfamily. Finally, we discuss the role that preexisting receptor subtypes may have had in the historic development of functionally distinct vasopressin and oxytocin lineages.
Cloning and Sequencing of LSCPR2-Poly(A) ϩ RNA was isolated from dissected vas deferens of L. stagnalis and reverse transcribed into oligo-(dT) 17  . The PCR mixture was overlaid with 70 l of mineral oil and incubated in a DNA thermal cycler (Perkin-Elmer Cetus) for 50 cycles of 1 min at 94°C, 1 min at 50°C, and 1 min at 72°C. Amplified cDNA fragments were digested with EcoRI and BamHI and separated on agarose gel, and fragments of the expected size were isolated, cloned into pBluescript (Stratagene), and sequenced. One fragment was used to screen several Lymnaea brain-specific ZAPII cDNA libraries in order to obtain a full-length clone. The insert-containing pBluescript phagemid corresponding to this clone (pBSCPR2) was rescued by in vivo excision and sequenced.
Functional Expression of LSCPR2 in Xenopus Oocytes-The proteinencoding region of pBSCPR2 was amplified using primers directed to the translation initiation and termination sequences of the cDNA. The primers contained at the 5Ј-end a recognition sequence for the restriction endonucleases BamHI and EcoRI, respectively. The sense primer contained the Kozak consensus sequence for transcription initiation (CCACCATG; Ref. 28) to replace the Lymnaea sequence at this position. Amplification was carried out in a DNA thermal cycler using 3 units of Ultma DNA polymerase (Perkin-Elmer Cetus), under conditions recommended by the supplier. Reactions were incubated for 12 cycles of 15 s at 94°C, 15 s at 60°C, and 1.5 min at 72°C. The cDNA fragment was digested with BamHI and EcoRI, ligated into pGEMHE (29), and 5Јcapped cRNA was transcribed in vitro from a 3Ј-end-linearized plasmid using 40 units of T7 RNA polymerase and Cap-Scribe transcription buffer (Boehringer Mannheim). After phenol extraction and ethanol precipitation, the cRNA was dissolved in bidistilled water, and 50 nl (ϳ3 ng RNA) was injected into Xenopus laevis oocytes. After incubating the injected oocytes for 2-3 days at 20°C (30), two-electrode voltageclamped oocytes were challenged with 1.5-min applications of Lysconopressin and various related peptides. Peptide-induced membrane currents were recorded at holding potentials of Ϫ50 mV. Dose-response curves were constructed by integrating the membrane currents over time and plotting these values (expressed as percentage of the maximal membrane current) semilogarithmically versus the ligand concentration. From the mean of three experiments the EC 50 value was calculated.
RT-PCR-Male and female sex organs and glands, kidney, skin, brain, salivary gland, and hepatopancreas were dissected, and total RNA was isolated from each tissue in triplicate. Equal amounts of RNA (3-5 g for each tissue) were reverse transcribed into oligo-(dT) 17primed cDNA using 200 units of Molony murine leukemia virus reverse transcriptase (Life Technologies, Inc.), and of each reaction 5% was used directly in a PCR experiment. Primers specific for either LSCPR1 or LSCPR2 were used to amplify conopressin receptor cDNAs, and primers specific for Lymnaea fructose 1,6-biphosphate aldolase 2 were used as a positive control for each cDNA preparation. After 30 cycles of amplification, PCR products were separated on agarose gel, transferred to Hybond N ϩ membrane (Amersham Corp.), hybridized to 5Ј-end 32 Plabeled primers specific for each product, and autoradiographed.
In Situ Hybridization-Digoxigenin-labeled run-off antisense RNA was synthesized from 5Ј-end-linearized plasmids pBSCPR1 (26) and pBSCPR2 using 20 units of T3 RNA polymerase and digoxigenin-UTP labeling mixture (both from Boehringer Mannheim). Hybridizations of alternately collected sections of Lymnaea brains were performed as described previously (26).
Phylogenetic Analysis-Sequences of the members of the vasopressin/oxytocin receptor family were extracted from a G protein-coupled receptor data base (31) and aligned using the program Clustal V (32). Alignments were adjusted manually, and the transmembrane domains, as defined by Baldwin (33), were utilized to generate a PAUP data matrix (34). An accepted mutation stepmatrix was calculated using the program AMP. 3 A branch and bound search was performed without the stepmatrix, and the three most parsimonious trees with tree lengths of 858 were selected. When the branch and bound search was repeated with the stepmatrix invoked, one of these trees was found to be the shortest. Using the PAUP program, 1,000 replicates of bootstrap analysis were performed using the heuristic search procedure. The calculated g 1 value of Ϫ0.869935 (p Ͼ 0.01) indicated that relevant phylogenetic information was contained within the data set (35).

RESULTS
Cloning of the LSCPR2 cDNA-PCR on vas deferens cDNA using primers TM6s and TM7as yielded four putative receptor cDNA fragments. Data base searches revealed that two fragments had significantly higher sequence similarity with the vertebrate vasopressin and oxytocin receptors than with any other G protein-coupled receptor. We recently reported the cloning of the full-length cDNA corresponding to the first fragment (26). Here, we used the second cDNA fragment to screen brain-specific ZAPII cDNA libraries for full-length clones. From 4 ϫ 10 6 independent clones that were screened, one clone hybridized and was isolated. The corresponding pBluescript phagemid, named pBSCPR2, was rescued by in vivo excision and sequenced from both strands. pBSCPR2 has an insert of 4.5 kilobase pairs and contains an open reading frame of 1,560 base pairs encoding a 520-amino acid protein with a predicted molecular mass of 59.4 kDa (Fig. 1A). The open reading frame is preceded by an in-frame stop codon at position Ϫ45, which indicates that the coding region is complete at the 5Ј-end.
Structural Characteristics of LSCPR2-Hydrophobicity analysis (36) of the predicted amino acid sequence revealed 7 hydrophobic domains characteristic for the 7 TMs of G proteincoupled receptors ( Fig. 1B; Ref. 33). Therefore, and because of its sequence similarity with vasopressin and oxytocin receptors and with LSCPR1, the predicted protein was tentatively named L. stagnalis conopressin receptor 2 (LSCPR2). LSCPR2 contains several potential sites for N-linked glycosylation and for phosphorylation (Fig. 1A). Consensus sequences for N-glycosylation (Asn-X-Ser/Thr; Ref. 37) occur predominantly in the extracellular N-terminal domain and may have a function in targeting the receptor protein to the plasma membrane (38). Consensus sequences for phosphorylation sites (39) include four sites for protein kinase C (Ser/Thr-X-Arg/Lys), two sites for casein kinase II (Ser/Thr-X-X-Asp/Glu), and three sites for cAMP-dependent protein kinase (Arg/Lys-Arg/Lys-X-Ser/Thr). These are restricted to the third intracellular loop (i.e. between TM5 and TM6) and the C terminus, suggesting that phosphorylation may be important in the modulation of G protein coupling and receptor function (1).
An amino acid sequence alignment of LSCPR2, LSCPR1, and receptors for vasopressin, oxytocin, and vasotocin from vertebrates is presented in Fig. 2. In the region from the beginning of TM1 to the end of TM7, LSCPR2 has 43% sequence identity with the human oxytocin receptor, the rat V1a, and the human V1b vasopressin receptors and the fish vasotocin receptor, and 39% sequence identity with LSCPR1 and the rat V2 vasopressin receptor. Sequence identity is highest in the transmembrane domains, especially in TM7, whereas the N-and Cterminal domains as well as the third intracellular loop show hardly any sequence identity.
Several amino acid residues (indicated by asterisks in Fig. 2) that are conserved only among the members of the vasopressin/ oxytocin receptor family are thought to be important in ligand binding (6). In LSCPR2, these residues are either identical or conserved, with the exception of a threonine residue replacing the glycine residue in the sequence Pro-Trp-Gly (just before TM5). Interestingly, the two conserved proline residues in the sequences Gly-Pro-Asp and Pro-Trp-Gly (just before TM3 and TM5, respectively), which are replaced by aspartate residues in LSCPR1, are unchanged in LSCPR2.
Two adjacent cysteines that are found 15 residues downstream of TM7 in the vasopressin and oxytocin receptors may be responsible for anchoring the C terminus to the plasma membrane through palmitoylation, as has been observed for the ␤-adrenergic receptor (40). In LSCPR2, there are 4 adjacent cysteines present 4 residues further downstream. These may be involved in linking the C terminus of the receptor to the plasma membrane through a palmitoylation anchor.
Functional Expression of LSCPR2 in Xenopus Oocytes-To demonstrate that LSCPR2 indeed is a receptor for Lysconopressin, we have expressed the LSCPR2 cDNA in Xenopus oocytes by injection of cRNA and studied the response of injected oocytes upon application of Lys-conopressin and related peptides. To improve cRNA stability and expression levels in Xenopus oocytes, the open reading frame of LSCPR2 was cloned in pGEMHE, between the 5Ј-and 3Ј-untranslated sequences of the Xenopus ␤-globin gene (29). Voltage-clamped oocytes that were previously injected with 5Ј-capped cRNA derived from this construct responded to application of conopressins by displaying inward membrane currents (Fig. 3A). Only application of Lys-conopressin and its synthetic oxytocinlike analog Ile-conopressin resulted in maximal responses. Argconopressin was about 10 times less potent, and all of the vertebrate peptides tested had no effects at concentrations as high as 10 Ϫ6 M. The effects of Lys-conopressin and Ile-conopressin were dose-dependent, with estimated EC 50 values of 86 and 96 nM, respectively (Fig. 3B). These values are about four times higher than observed for LSCPR1 (26).
Differential Expression of the LSCPR1 and LSCPR2 Genes-Since LSCPR1 and LSCPR2 mRNA levels are undetectable using Northern blot analysis (data not shown), we applied RT-PCR to determine possible co-expression of the LSCPR1 and LSCPR2 genes in peripheral tissues and brain. The expression of both receptor genes is restricted to the brain and the reproductive organs (Fig. 4). Interestingly, the anterior part of the vas deferens contains exclusively LSCPR1 mRNA, whereas the spermoviduct contains exclusively LSCPR2 mRNA. In the posterior part of the vas deferens, a small amount of both transcripts was detected. In situ hybridization on alternate sections of the brain revealed that the two receptor genes are expressed in mutually exclusive sets of neurons. As reported previously (26), the LSCPR1 gene is expressed in neurons in the anterior lobe of the right cerebral ganglion and the pedal Ib cluster, in a subpopulation of the neuroendocrine light green cells, and in some unidentified neurons throughout the brain (not shown). The LSCPR2 gene is predominantly expressed in neurons in the visceral ganglion (Fig. 5A) and the right parietal ganglion (Fig. 5B) and in some neurons that are located in close proximity of the anterior lobes of the cerebral ganglia (Fig. 5C). Control hybridizations with sense cRNA probes were all negative.
Phylogenetic Relationships of Lys-Conopressin Receptors and Their Vertebrate Counterparts-From a sequence alignment of the transmembrane domains of the members of the vasopressin/oxytocin receptor family, a most parsimonious phylogenetic tree was calculated. This tree (Fig. 6) has an unweighted tree . Amino acid residues that are identical in all receptor proteins are shown in boldface. Asterisks indicate amino acid residues that are highly conserved in the vertebrate vasopressin/oxytocin receptor family, but not in other G protein-coupled receptors and have been suggested to be important in receptor-ligand interaction (6). Bars indicate the seven putative transmembrane domains.
length of 858 and is the best tree found in over 327,128 trees sampled. It demonstrates that LSCPR1 and LSCPR2 are clearly related to the vasopressin/oxytocin receptor family of the vertebrates. Bootstrap analysis of the tree and the data set show that the assignment of LSCPR1 and LSCPR2 to the vasopressin/oxytocin receptor clade is supported with over 96% confidence. These findings strongly suggest that the conopres-sin receptors and their vertebrate counterparts evolved from a common ancestral receptor. Moreover, the tree shows that LSCPR2 is more closely related to the vertebrate receptors than LSCPR1 and that the separate evolutionary history of LSCPR1 and LSCPR2 is probably much longer than that of the vertebrate receptor subtypes.

DISCUSSION
Structure and Functional Expression of LSCPR2-To provide molecular data for a theory of the co-evolution of peptidereceptor pairs, we have cloned various Lys-conopressin receptors in the mollusc L. stagnalis. Using degenerate oligonucleotides directed to conserved sequences in the mammalian vasopressin and oxytocin receptors, we amplified similar sequences from the vas deferens of Lymnaea. The cloning of the full-length cDNA corresponding to one of these amplification products was recently reported (26). Here, we have used a second amplification product exhibiting high sequence identity with vertebrate vasopressin and oxytocin receptors in order to obtain a full-length cDNA. Hydrophobicity analysis of the predicted protein sequence of this clone, named LSCPR2, revealed that it contains seven hydrophobic regions similar to the seven transmembrane domains found in all members of the G protein-coupled receptor superfamily. In addition, LSCPR2 contains amino acid residues in the first and second extracellular loops that are characteristic for the members of the vasopres-  14). The LSCPR1 transcript could be amplified from the brain and the anterior and posterior vas deferens, and the LSCPR2 transcript could be amplified from the brain, the spermoviduct and the posterior vas deferens. Lanes 15 and 16 contain PCR products derived from the LSCPR1 and LSCPR2 clones, respectively, and serve as positive controls for the PCR and the hybridization procedure. sin/oxytocin receptor family (Fig. 2). LSCPR2 is activated by Lys-conopressin at physiological concentrations, but not by other naturally occurring vasopressin-or oxytocin-related peptides (Fig. 3). Together, these data indicate that LSCPR2 is a functional Lys-conopressin receptor, and that LSCPR1 and LSCPR2 represent distinct subtypes of Lys-conopressin receptors in Lymnaea.
Compared with the vertebrate receptors, the LSCPR2 protein has the highest sequence identity with the mammalian V1a and V1b vasopressin receptors (17,19), the mammalian oxytocin receptor (20), and the vasotocin receptor from teleost fish (41), which all couple to the inositol trisphosphate/Ca 2ϩ signaling pathway. Sequence identity with the mammalian V2 vasopressin receptor (13,14), which is coupled to adenylate cyclase, is less. Furthermore, Xenopus oocytes expressing the LSCPR2 protein respond to application of Lys-conopressin by displaying a dose-dependent inward chloride current (Fig. 3), a response that is characteristic for receptors that activate the inositol trisphosphate/Ca 2ϩ signal transduction pathway (30). These findings indicate that most likely LSCPR2 is coupled to a G protein that activates the inositol trisphosphate/Ca 2ϩ signal transduction pathway. LSCPR1 has similar signal transduction characteristics, however, half-maximal receptor activation is reached at four times lower concentrations of Lysconopressin (26). The first and second extracellular loops of the vasopressin and oxytocin receptors have been suggested to form a ligand binding pocket (6). Interestingly, the corresponding loops in LSCPR1 contain two unique aspartate residues that are absent in LSCPR2 (Fig. 2). Since conopressins differ from other vasopressin-and oxytocin-related peptides in that they have a negatively charged arginine residue at position 4 (22), we have suggested that either of these positively charged aspartate residues may be important in the interaction of LSCPR1 with this arginine residue (26). The absence of the aspartate residues in LSCPR2 might explain the fact that this receptor is less sensitive to Lys-conopressin than LSCPR1.
Differential Expression of Conopressin Receptor Subtypes in Lymnaea-To investigate the physiological significance of multiple conopressin receptors in Lymnaea, we compared the peripheral and neuronal distribution of LSCPR1 and LSCPR2 by means of RT-PCR and in situ hybridization. The expression of both receptors is restricted to the reproductive organs and the brain. LSCPR1 is exclusively expressed in the vas deferens ( Fig. 4) and is involved in the control of muscular activity during transport of semen (26). LSCPR2 is predominantly expressed in the spermoviduct, and a small amount of transcript could be detected in the posterior part of the vas deferens (Fig.  4). In the spermoviduct, both eggs and sperm are transported, and fertilization of eggs and resorption of unused sperm occur. Interestingly, Lys-conopressin mRNA could also be detected in this tissue, 4 suggesting that LSCPR2 might mediate auto-or paracrine actions of Lys-conopressin in the control of any of the reproductive functions of the spermoviduct. Similarly, vasopressin has been suggested to have an autocrine regulatory role in Leydig cells of rodent testis (42). Although LSCPR1 and LSCPR2 transcripts are co-localized in the posterior part of the vas deferens, the LSCPR1 mRNA detected here is probably synthesized in central neurons that innervate the vas deferens (26). The cellular source of LSCPR2 mRNA in the vas deferens remains unclear. In the brain, co-localization of LSCPR1 and LSCPR2 transcripts could not be observed.
The differential distribution of the two Lys-conopressin receptors suggests that they must have different functions. In the periphery, they are most likely involved in the regulation of distinct aspects of reproduction, whereas in the brain, they 4 R. E. van Kesteren, unpublished results. probably mediate transmitter-like or modulatory effects of Lysconopressin on different types of neurons. A possible explanation for the existence of multiple conopressin receptors is that it allows a stimulus-dependent differential expression of the receptor genes in various target tissues and cells, thus providing a mechanism for a spatio-temporal co-ordination of Lysconopressin actions in otherwise conflicting types of behavior and physiological processes. Similarly, transcriptional regulation of the human oxytocin receptor gene plays an important role in the physiologically relevant increase in the number of oxytocin receptors in the uterus of pregnant females at the onset of labor (20). In addition, the difference in the sensitivity of the two receptors to Lys-conopressin that we observed may be physiologically relevant in this respect as well.
Prospects for Understanding the Co-evolution of Specifically Interacting Peptide-Receptor Pairs-Peptides are extensively employed as messenger molecules in the various communication systems that are involved in the control of physiological processes and behavior (43). These systems, which include wiring transmission (communication via the synapse), volume transmission (communication via the intercellular space), hormonal transmission (communication via the circulation), and environmental transmission, emerged at various stages during evolution. Due to differences in requirements for optimized information exchange, both peptide design and specificity of peptide-receptor interactions are under different selection pressures in the various communication channels (43,44). With the evolution of complex organisms such as vertebrates, many peptides and receptors were recruited into the hormonal communication channel, thus increasing the demands for specificity in order to exclude cross-activation of the many different receptors that can be reached. In view of the great variety in both size and conformational flexibility of peptides, it seems plausible that different structural solutions emerged in response to the need for specificity of peptide-receptor interactions. Large and flexible peptides such as luteinizing hormone and follicle-stimulating hormone, which like vasopressin and oxytocin evolved from a single ancestral gene, have acquired distinct regions of many amino acid residues that exclude cross-activation of receptors (45). Consequently, these specificity determinants are mirrored in the luteinizing hormone and follicle-stimulating hormone receptors at entirely different locations.
As we have shown, specificity determinants in the small and conformationally constrainted peptides of the vasopressin/oxytocin superfamily evolved in a different way. Residues 1-6 of these peptides are important for high affinity binding to the receptor (26), whereas specific receptor activation depends on the chemical nature of the single amino acid residue at position 8. In contrast to the development of luteinizing hormone and follicle-stimulating hormone and their receptors, where separation of large specificity domains in both the peptides and the receptors allowed for a continuous refinement of specificity during evolution (45), functionally distinct vasopressin and oxytocin peptides can only have evolved by trial-and-error mutation of residue 8 in the presence of a nondiscriminative receptor such as LSCPR2. This trial-and-error phase may be reflected in various present day species of cartilaginous fish. Unlike the evolutionary stable vasopressin lineage, oxytocinlike peptides are very diverse in this primitive group of the vertebrates (7). To none of these peptides has a function been assigned, and they have very poor uterotonic activity in mammals (46). Hence, it will be of interest to examine whether functional receptors exist for these peptides. If not, they must be considered relicts of an early step in the historic development of the oxytocin lineage of bioactive peptides.
To further demonstrate that a nondiscriminative LSCPR2like receptor may have been ancestral to the vasopressin/oxytocin receptor family, we studied the phylogenetic relationships of the Lys-conopressin receptors and their vertebrate counterparts using parsimony analysis. Phylogenetic reconstruction is nowadays accepted as a valid method to study the evolutionary histories of families of related receptors and provides useful insights into the structures and functions of the individual members (47). 5 Our data indicate that LSCPR2 is indeed more closely related to the vertebrate vasopressin and oxytocin receptors than LSCPR1 (Fig. 6). Therefore, LSCPR2 most likely is a present day representative of the ancestral receptor to the vertebrate receptors. In addition, the phylogenetic analysis shows that LSCPR1 and LSCPR2 probably result from an ancient receptor duplication that occurred before separate receptor types in the vertebrates evolved. Thus, multiple receptors may have existed before separate lineages of vasopressinand oxytocin-related peptides evolved. Since the evolution of functionally distinct peptide lineages from a common ligandreceptor pair requires not only the introduction of specificity determinants on both the peptides and the receptors but also a differential cellular pattern of expression of distinct receptor types, we suggest that preexistence of differentially expressed receptor subtypes was of significant importance in the functional divergence of vasopressin and oxytocin in the vertebrates.
As in the peptides of the vasopressin/oxytocin superfamily, specificity determinants in the corresponding receptors are probably restricted to small parts of the receptor molecules (6). LSCPR2 can be very useful in the search for these determinants. Experiments involving chimeric receptors and swapping of putative binding domains might mimmick the evolutionary process of introducing specificity determinants in a nondiscriminative ancestral receptor such as LSCPR2, and increase our understanding of the molecular basis of peptide-receptor interactions. This then might enable the rational design of highly potent and specific receptor agonists and antagonists.