Nephrogenic Diabetes Insipidus A V2 VASOPRESSIN RECEPTOR UNABLE TO STIMULATE ADENYLYL CYCLASE*

The coding region of the human vasopressin type 2 receptor gene bears mutations in the individuals affected with congenital nephrogenic diabetes insipi- dus, a disease characterized by the inability of the kidney to concentrate urine in response to vaso- pressin. Although it is assumed that the mutations result in loss of receptor function, proof of this hy-pothesis is lacking. We introduced one of these natu- rally occurring point mutations leading to a single amino acid change (Arg’” + His) into wild type cDNA. The mutant protein was expressed, and the functional properties of the receptor were examined. The mutant receptor exhibited an unaltered binding affinity for vasopressin compared to the wild type but failed to stimulate the GJadenylyl cyclase sys-tem. These data provide biochemical proof that the mutant receptor is the cause of the disease.

death. The disease shows a sex-linked recessive inheritance and was mapped to the q28-qter portion of the human X chromosome by genetic linkage (Kambouris et al., 1988;Knoers et al., 1988). Cloning of the human V2R cDNA (Birnbaumer et al., 1992b) enabled us to map the V2R gene to the vicinity of the CNDI locus (Seibold et al., 1992). We also showed that CNDI patients have mutations in the V2R gene (Rosenthal et al., 1992). To date, 16 distinct Grant DK 21-244 (to M. B.1. The costs of publication of this article were * This work was supported in part by National Institutes of Health defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $Recipient of a Heisenberg Fellowship from the Deutsche Forschungsgemeinschaft. V2R mutations occurring in patients with CNDI have been reported (Pan et al., 1992;van den Ouweland et al., 1992;Davies, 1992;Bichet et al., 1993). Six mutations lead to truncated forms of the receptor protein; none of these mutant receptors is expected to have biological activity. Ten mutations result in substitution of single amino acids, of which six are predicted to modify the extracellular portion of the receptor, three in transmembrane segments, and one at the junction of the third transmembrane segment and the second intracellular loop of the receptor. Although it is assumed that these single amino acid changes lead to receptor inactivation, biochemical proof of receptor inactivation has not yet been obtained. In the present study we concentrated on the last of the mutations mentioned above referred to as Q2 (Bichet et al., 1992(Bichet et al., , 1993 and sought to determine whether it indeed impairs receptor function, and if so, which aspect of receptor function would be affected. The Q2 mutation is a G + A transition in codon 137 of the V2R resulting in the change of an arginine that is highly conserved in G protein-coupled receptors to histidine (R137H; Fig. 1). Because of the location near the plasma membrane/ cytoplasm interface we hypothesized that the Q2 V2R mutant might have normal or close to normal ligand binding but impaired signaling activity. We report here that the mutant receptor binds AVP with an affinity very similar to that of the wild type receptor but fails to stimulate adenylyl cyclase.

Radiochemi~als-[~H]
Aginine vasopressin, specific activity 60 Ci/ mmol, was purchased from Du Pont-New England Nuclear; [a-3*P]ATP was from the NICHD Center for Population Research and Studies in Reproductive Biology at Baylor College of Medicine; [3Hlcyclic 3'3'-AMP was from ICN Biochemicals, Irvine, CA.
Construction ofphV2R-Q2"The human cDNA encoding the human V2R (Birnbaumer et al., 1992b) was excised from the Bluescript plasmid with EcoRI and cloned into the replicative form of M13mp18. Sitedirected mutagenesis was performed following the method of Nakamaye and Eckstein (1986) using reagents purchased as a kit from Amersham Corp. An antisense oligonucleotide of 27 bases that contained an A instead of G at the position corresponding to nucleotide 410 of the open reading frame was annealed to the single-stranded cDNA. Following in vitro mutagenesis, the complete cDNA was sequenced by the method of Sanger et al. (1977) to verify that only the desired mutation had been introduced. The modified cDNA (Q2 cDNA) was excised with EcoRI, and the "sticky" ends were filled with the Klenow fragment of DNA polymerase I and ligated into the expression vector pKNH (Birnbaumer et al., 1992b).
Dansient Expression in COS.M6 Cells-COS.M6 cells were grown in DMEM (high glucose), supplemented with 10% heat-inactivated FBS, penicillin (50 unitdml) and streptomycin (50 pg/ml). Subconfluent cells were plated at a density of 1.0-2.0 X lo6 cells/lOO-mm dish and transfected by a modification of the DEAE-dextran method of Luthman and Magnusson (1983) with 3 pg of plasmid DNNplate. Thereafter cells were kept 4 h at 37 "C, exposed to 10% dimethyl sulfoxide in HBSS for 3 min, and then to 100 p~ chloroquine in DMEM containing 2% FBS for 3-4 h at 37 "C. For binding assays cells were trypsinized 24 h after transfection, seeded at a density of 0.5-1.0 X lo6 celldwell in 12-well plates, and assayed after another 24 h. For adenylyl cyclase assays the cells were incubated for 48 h, harvested with a rubber policeman, and homogenized with a Dounce homogenizer. The homogenate was then assayed for adenylyl cyclase activity.
Hormone Binding to Intact Cells-Cells were plated in 12-well plates a t a density of 1.0 X lo6 cells/well. Binding assays were performed the following day as described (Birnbaumer et al., 1992a).

Adenylyl Cyclase Activity in Cell
Homogenates-Adenylyl cyclase activity was assayed as described (Birnbaumer et al., 1992a). amino acid sequence of the wild type V2R is presented, and the site of the Q2 mutation is indicated. The putative transmembrane a-helices are shown as suggested by Kyte and Doolitle (1982) analysis of the protein sequence; a disulfide bridge is postulated between the first and second extracellular loops based on mutational studies with rhodopsin (Karnik et al., 1988) and the conservation of cognate cysteines in all G protein-coupled receptors. Folding of NH2 and COOH termini and of longer intracellular loops is arbitrary. Anchoring of the carboxyl-terminal portion by fatty acids is assumed to be analogous to that of rhodopsin (two palmitoyl anchors). Amino acids are given in the single-letter code.  The results are summarized in Table I. Receptors were present on the surface of cells bearing the mutant as well as the wild type cDNA. The number of receptors per cell was approximately 14 times lower for the Q2 mutant than for the wild type receptor. In contrast, the KO values for A W deduced from saturation binding experiment (not shown), did not differ significantly (6.7 and 10.4 nM for the Q2 and wild type receptor, respectively). Stimulation of adenylyl cyclase activity by the endogenous adrenergic receptor was similar in cells bearing the wild type or the mutant cDNA (Table I). However, only cells transfected with wild type cDNA showed an increase in adenylyl cyclase activity when challenged with 100 nM A". The stimulation of adenylyl cyclase activity measured in COS.M6 cells after transient expression is only 2-%fold over basal. Thus, it was possible that the assay failed to detect a weak stimulation of adenylyl cyclase activity by the reduced number of Q2 receptors. Since the V2R expressed in stably transfected L cells shows a robust stimulation of adenylyl cyclase (Birnbaumer et al., 1990b), we decided to develop L cells expressing the Q2 mutant receptor to analyze the correlation between receptor abundance and stimulation of adenylyl cyclase activity.

Dansient Expression in
Stable Expression in Ltk-Cells-Since the mutant receptor protein is successfully transported to the plasma membrane in COS.M6 cells the stable transformants expressing the receptor were identified by binding assays (Liao et al., 1988). G418resistant clonal cells transfected with the mutant cDNA were plated in duplicate in 12-well plates and single point binding assays were performed: one for total binding and one for nonspecific binding in the presence of 10 PM unlabeled AW. We identified three cell lines expressing the Q2 mutation that were expanded and cloned by limiting dilution. The Q2-3 cell line was chosen for further experimentation because it expressed the highest number of mutant receptors (see below); the other lines contained 8,000 (Q2-6) and 15,000 (Q2-24) sites per cell.
Saturation binding of L3H1AVP to Q2-3 cells and to HTB-2 cells (expressing the wild type receptor), was determined in the same experiment (Fig. 2). The Q2 receptor exhibited an affinity for vasopressin (KD= 8.4 2 0.5 nM) that was comparable to that of the wild type receptor (KD= 6.4 2 0.2 nM) as determined in three experiments; values were similar to those observed in COS cells. The Q2-3 cell line expressed 25,000 to 30,000 receptors per cell. This value is close to the number of wild type receptors present on the surface of the HTB-2 cells (15,000-20,000 sites per cell).
Because of the similar receptor abundance of Q2-3 and HTB-2 cells (Birnbaumer et al., 1992b), we compared the adenylyl cyclase response to vasopressin in these two cell lines. As illustrated in Fig. 3, the mutant receptors were unable to trigger adenylyl cyclase stimulation even at very high concentrations of hormone, at which the receptor is fully occupied. In the same assay, the adenylyl cyclase activity of the HTB-2 cells was stimulated 15-fold. The responses of adenylyl cyclase to forskolin and prostaglandin El (acting through endogenous prostaglandin receptors), were similar in both cells. Similar results were obtained with the Q2-6 and Q2-24 cells (data not shown). It is worth mentioning that the adenylyl cyclase activity of the LV2.E2 cell, an L cell expressing only 5,000 receptors per cell (Birnbaumer et al., 1992b), is stimulated 7-fold by AVP. The present data identifies the biochemical defect of the Q2 mutant receptor as failure to activate G, rather than an alteration of the affinity for AVP.

DISCUSSION
The present report proves that the Q2 mutation is indeed responsible for the CNDI phenotype. Although the mutant receptor shows normal binding properties, it is unable to stimulate the GJadenylyl cyclase system. We also observed in the COS cells a remarkable difference in the expression of the wild type uersus the Q2 receptor. The low number of mutant receptors could be due to decreased stability of the mutant RNA, although we consider this unlikely. It seems more likely that the Arg + His change at codon 137 interferes with proper folding of the nascent protein giving rise to a "mis-folded receptor, which is less efficiently translocated across the endo- plasmic reticulum and the Golgi apparatus. The modification of protein structure may also interfere with appropriate insertion into the plasma membrane or result in an increased degradation rate in the endoplasmic reticulum or Golgi cisternae. Because the mutant receptor has retained AVP binding activity with an affinity similar to that of the wild type, we assume that protein half-life rather than its insertion into the plasma membrane is altered. In either case we expect the affected members of the Q2 families to have reduced number of receptor protein on the surface of the corresponding kidney cells. Regardless of these considerations we expect the Q2 receptor to have a cytoplasmic conformation different from that of the wild type, and as a consequence, to lack the ability to contact G. or to activate the GDP/GTP exchange. The identification of the stably transfected cells expressing the mutant receptor was done by binding, a less sensitive assay than the measurement of AVP-responsive adenylyl cyclase that we have used for the identification of cells expressing the wild type receptor (Birnbaumer et al., 1992b). Thus, the screening procedure selected for cells with a high number of receptors. The majority of the G418-resistant clones expressed the Q2 mutant receptor at much lower numbers than the three cell lines reported here. The Q2 mutation is of interest not only because of its localization and biochemical consequences but also because it alters an Asp/Glu-Arg motif conserved in many G protein-coupled receptors (Savarese and Fraser, 19921, including, among others, neurotransmitter receptors, peptide hormone receptors, glycoprotein receptors, and light and odor receptors. Six receptors lacking this double amino acid motif form a subfamily including the receptors for parathyroid hormone, calcitonin, and the glucagon-related peptides glucagon, glucagon-like peptide I, vasoactive intestinal peptide, and secretin (Thorens, 1992;Jelinek et al., 1993).
This motif has received attention already in two G protein-  (1988) showed an unaltered EC50 value for stimulation of adenylyl cyclase by (-)isoproterenol, and a presumably unaltered affinity for agonist. However there was a 50% reduction in the maximal stimulation of the G,/adenylyl cyclase system, probably reflecting a reduced ability of the mutant receptor to couple to G,. The [Asp130 -+ AsnIpAR mutant (Fraser et al., 1988) had an increased affinity for agonist and retained some of the effects of GTP on agonist binding, but it had totally lost the ability to stimulate the G,/adenylyl cyclase system. The data indicated that the mutant interacts with G, ( i e . it forms a ternary complex) but does not promote nucleotide exchange on the G, a-subunit. Working with bovine rhodopsin Khorana and collaborators constructed a [ G l~'~~-A r g l~~ -+ Arg-Glulopsin. The mutant retained its ability to bind retinal and supported light-induced retinal isomerization but failed to interact with transducin, as was evident from the failure to be stabilized as metarhodopsin-I1 by transducin, akin to the induction of the high affinity state of the p-AR, and from the failure to stimulate the GTPase activity of transducin (Franke et al., 1990). The same results were obtained with [Arg135 -+ Glnlopsin; retinal bound normally, but interaction with transducin did not occur (Franke et al., 1992).
In two cases of individuals with autosomal dominant retinitis pigmentosa, Sung et al. (1991a) identified changes in the codon of human opsin. The naturally occurring mutants, AI-^'^^ + Leulopsin and -+ Trplopsin did not bind retinal (Sung et al., 1991b), in contrast to the bovine opsin mutants mentioned above.