Functional Rescue of the Nephrogenic Diabetes Insipidus-causing Vasopressin V2 Receptor Mutants G185C and R202C by a Second Site Suppressor Mutation*

Mutations in the gene of the G protein-coupled vasopressin V2 receptor (V2 receptor) cause X-linked nephrogenic diabetes insipidus (NDI). Most of the missense mutations on the extracellular face of the receptor introduce additional cysteine residues. Several groups have proposed that these residues might disrupt the conserved disulfide bond of the V2 receptor. To test this hypothesis, we first calculated a structure model of the extracellular receptor domains. The model suggests that the additional cysteine residues may form a second disulfide bond with the free, nonconserved extracellular cysteine residue Cys-195 rather than impairing the conserved bond. To address this question experimentally, we used the NDI-causing mutant receptors G185C and R202C. Their Cys-195 residues were replaced by alanine to eliminate the hypothetical second disulfide bonds. This second site mutation led to functional rescue of both NDI-causing mutant receptors, strongly suggesting that the second disulfide bonds are indeed formed. Furthermore we show that residue Cys-195, which is sensitive to “additional cysteine” mutations, is not conserved among the V2 receptors of other species and that the presence of an uneven number of extracellular cysteine residues, as in the human V2 receptor, is rare among class I G protein-coupled receptors.

Mutations in the gene of the G protein-coupled vasopressin V 2 receptor (V 2 receptor) cause X-linked nephrogenic diabetes insipidus (NDI). Most of the missense mutations on the extracellular face of the receptor introduce additional cysteine residues. Several groups have proposed that these residues might disrupt the conserved disulfide bond of the V 2 receptor. To test this hypothesis, we first calculated a structure model of the extracellular receptor domains. The model suggests that the additional cysteine residues may form a second disulfide bond with the free, nonconserved extracellular cysteine residue Cys-195 rather than impairing the conserved bond. To address this question experimentally, we used the NDI-causing mutant receptors G185C and R202C. Their Cys-195 residues were replaced by alanine to eliminate the hypothetical second disulfide bonds. This second site mutation led to functional rescue of both NDI-causing mutant receptors, strongly suggesting that the second disulfide bonds are indeed formed. Furthermore we show that residue Cys-195, which is sensitive to "additional cysteine" mutations, is not conserved among the V 2 receptors of other species and that the presence of an uneven number of extracellular cysteine residues, as in the human V 2 receptor, is rare among class I G protein-coupled receptors.
The vasopressin V 2 receptor (V 2 receptor) 1 belongs to the large family of G protein-coupled receptors (GPCRs). The receptor is expressed in the basolateral membrane of the principal epithelial cells of the renal collecting duct, where it mediates the antidiuretic action of the hormone AVP (for review, see Ref. 1). Activation of the V 2 receptor leads to stimulation of the G s /adenylyl cyclase system. The subsequent rise in intracellular cAMP induces the fusion of vesicles containing water channels (aquaporin 2) with the apical membrane which is thus rendered water permeable. As a consequence, water is reabsorbed from the lumen of the collecting duct.
Mutations in the gene of the V 2 receptor are the cause of X-linked NDI (for review, see Ref. 2), a disease characterized by the inability of the kidney to concentrate urine despite normal or elevated levels of AVP. More than 170 V 2 receptor mutations have been documented, approximately half of which are missense mutations. Most of the missense mutations are clustered within the transmembrane domains of the receptor. These mutations usually affect receptor folding in the bilayer and lead to intracellular retention of the misfolded forms by the quality control system of the endoplasmic reticulum. Compared with the numerous mutations found in the transmembrane domains, only few mutations are located on the cytoplasmic or extracellular faces of the receptor. It is striking that most of the extracellular mutations introduce additional cysteine residues. Such mutations include R106C (3), R181C (4,5), G185C (6), R202C (3,6,7), and Y205C (6, 8,9). The most obvious explanation for the origin of the defect caused by these additional extracellular cysteine residues is that they impair formation of the single disulfide bond of the V 2 receptor which connects residue Cys-112 of the first with Cys-192 of the second extracellular loop and which is conserved in the GPCR family. Several groups have proposed that the mutated residues form an alternative bond with either one of the two conserved cysteine residues (5,6,9,10). Disease-causing mutations that lead to the introduction of additional cysteine residues have been observed for other GPCRs. The Y178C mutation of rhodopsin was shown to cause retinitis pigmentosa and is suspected to impair the formation of the conserved disulfide bond of this receptor (11). The analysis of the molecular basis of the defect caused by these "additional cysteine" mutations of GPCRs is thus of general interest.
In the case of the V 2 receptor, recent studies called into question whether the additional cysteine residues impaired formation of the conserved disulfide bond (12); mutation of the conserved cysteine residues led to nonfunctional receptors with a strong transport defect, whereas the NDI-causing mutations G185C and R202C led to binding-defective but transport-competent receptors. If the additional cysteines of the NDI-causing mutant receptors were to disrupt the conserved disulfide bond, an impaired intracellular transport of these mutant receptors would also have been expected.
Here we have analyzed in detail the defects of the NDIcausing mutant receptors G185C and R202C. We present molecular modeling and mutagenesis data that strongly suggest that the additional cysteine residues of both mutant receptors participate in a second disulfide bond with residue Cys-195 of the V 2 receptor rather than impairing formation of the conserved bond.
DNA Manipulations-Standard DNA preparations and manipulations were carried out according to the handbook of Sambrook and co-workers (14). Nucleotide sequences of DNA constructs were verified using the FS dye terminator kit from PerkinElmer Life Sciences.
Construction of NDI-causing Mutant Receptors Carrying an Additional C195A Mutation-Plasmid pWT.GFP, encoding a fusion of the red-shifted variant of GFP to residue Lys-367 of the V 2 receptor (i.e. to the entire receptor except for the four C-terminal residues) has been described previously (15). The C195A mutation was introduced into pWT.GFP by the use of the Quick Change site-directed mutagenesis kit. A primer with the sequence 5Ј-CTGACTGCTGGGCCGCCTTTGCG-GAGCCC-3Ј and its complementary equivalent were employed. The resulting plasmid C195A.GFP was then used to introduce the NDIcausing mutations G185C and R202C. Primers with the sequences 5Ј-CAGCGCAACGTGGAATGCGGCAGCGGGGTCAC-3Ј (resulting plasmid pG185C/C195A.GFP) and 5Ј-CGGAGCCCTGGGGCTGTCG-CACCTATGTC-3Ј (resulting plasmid pR202C/C195A.GFP) and their complementary equivalents were used.
Cell Culture and Transfection Methods-All experiments in this study were carried out with transiently transfected HEK 293 cells.
Cells were cultured on poly-L-lysine-coated material in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin in a 5% CO 2 atmosphere. Cells were transfected with LipofectAMINE according to the supplier's recommendations.
For the [ 3 H]AVP binding assay, 5 ϫ 10 4 cells in a 15-mm-diameter well (of a 24-well plate) were transfected with 250 ng of plasmid DNA and 2 l of LipofectAMINE. The cell surface biotinylation assay and the isolation of crude membranes for adenylyl cyclase assays and for immunoblots proceeded from confluent cells grown on 60-mm-diameter dishes. Here, 5 ϫ 10 5 cells were transfected with 3.5 g of plasmid DNA and 26.25 l of LipofectAMINE. The transfection for laser scanning microscopy was carried out with cells grown in 35-mm-diameter dishes containing glass coverslips. Cells (4 ϫ 10 4 ) were transfected with 1 g of plasmid DNA and 7.5 l of LipofectAMINE. In all cases, cells were incubated further for 48 h after removal of the transfection reagent.
[ 3 H]AVP Binding Assay and Adenylyl Cyclase Assay-The [ 3 H]AVP binding assay was carried out with intact, transiently transfected HEK 293 cells as described previously for African green monkey kidney (COS.M6) cells (16). However, HEK 293 cells were grown in 24-well plates (15-mm diameter/well) rather than the 35-mm-diameter dishes described for COS.M6 cells. The adenylyl cyclase assay was carried out with nuclei-free crude membranes of transiently transfected HEK 293 cells as described previously for stably transfected L tkϪ cells (16).
Glycosylation State Analysis-Crude membranes of transiently transfected HEK 293 cells expressing the GFP-tagged V 2 receptors were isolated from confluent cells grown on 60-mm-diameter dishes as described previously for COS.M6 cells (17). Membrane proteins were incubated with or without endoglycosidase H or PNGase F. Receptors were detected by immunoblotting using a rabbit anti-GFP antiserum and 125 I-conjugated anti-rabbit IgG as described previously (13).
Cell Surface Biotinylation Assay-Transiently transfected HEK 293 cells expressing the GFP-tagged V 2 receptors were grown in a 60-mmdiameter dish to confluence. Cells were washed three times with icecold PBS-CM buffer (PBS containing 0.1 mM CaCl 2 and 1 mM MgCl 2 , pH 7.4). Cell surface proteins were labeled by incubating cells with PBS-CM containing 1 mg/ml Sulfo-NHS-Biotin for 30 min at 4°C. Labeling reactions were quenched by replacing the biotin solution with 1 ml of NH 4 Cl solution (50 mM in PBS-CM). After 10 min, the cells were washed three times with ice-cold PBS-CM. 1 ml of ice-cold lysis buffer (1% Triton X-100, 0.1% SDS, 50 mM Tris-HCl, 150 mM NaCl, 1 mM Na-EDTA, pH 8.0) was added to the dishes, and membrane proteins were solubilized for 1 h at 4°C. Insoluble debris was removed by centrifugation (20 min, 4°C, 47,000 ϫ g), and biotinylated proteins were recovered from the supernatant by a 1.5-h incubation at 4°C with NeutrAvidin-agarose beads. Beads were sedimented (3 min, 17,000 ϫ g, 4°C), washed twice with buffer (0.5% Triton X-100, 0.1% SDS, 50 mM FIG. 1. Assignment of the retrieved protein fragments from the Brookhaven 3D protein data base to the extracellular domains of the V 2 receptor. The overlapping sequences are aligned to a partial sequence of the Nterminal tail (Ntt) and to ECL1, ECL2, and ECL3. The V 2 receptor sequences are indicated by a black frame. Identical residues of the fragments are indicated by uppercase letters; similar and nonsimilar residues are in lowercase letters. The data base accession numbers of the proteins containing the respective fragments sequences are shown on the right. The presence of more than one accession number indicates the occurrence of similar fragments with comparable conformations in different proteins. Only the fragments from the protein represented by the first accession numbers were used for the modeling procedure in these cases.
Visualization of GFP-tagged Receptors by Confocal Laser Scanning Microscopy-Transiently transfected HEK 293 cells expressing the GFP-tagged receptors were grown on glass coverslips. Cells were washed twice with PBS (pH 7.4), transferred immediately into a selfmade chamber (details on request), and covered with 1 ml PBS (pH 7.4). The GFP fluorescence signals were visualized on a Zeiss 410 invert laser scanning microscope ( exc ϭ 488 nm, em ϭ Ͼ515 nm). Subsequently, the cell surface of the same cells was stained with 0.05% trypan blue as described (15). Trypan blue fluorescence ( exc ϭ 543 nm, em ϭ Ͼ590 nm) was recorded on a second channel, and its overlap with the GFP signals was computed.
V 2 Receptor Model Building-A structure model of the transmembrane and the intracellular domains of the V 2 receptor was described previously (13). For this model, packing of the transmembrane helices was based on electron density maps of frog rhodopsin (18).
The structure model of the extracellular domains was assembled stepwise taking this previously described model as a foundation. The starting conformations of the extracellular domains were obtained by assembling the known conformations of identical or similar fragments of other proteins (3-10 residues) retrieved from the Brookhaven 3D protein data bank (Brookhaven National Laboratory, Upton, NY). Overlapping fragments with comparable conformations in several different proteins, indicating a common conformational propensity, are listed in Fig. 1 with their data base identification number. The strategy of assembling thoroughly selected fragments with known conformations has the advantage over loop search algorithms in that assignment of the orientation of the hydrophobic and hydrophilic side chains to the hydrophobic and to the water phase respectively is finer. Model components were assembled with the biopolymer module of the Sybyl program package (TRIPOS Inc. St Louis, MO) and minimized by an AMBER 5.0 force field. The stability of the resulting receptor model was finally assessed as described (19). Molecular dynamics simulations maintaining helix stability only by backbone H-bond constraints were performed at 300 K for 200 ps using AMBER 5.0 force field conditions in vacuo. The overall backbone conformations of the extracellular domains remained stable during the molecular dynamics runs, indicating a stable starting conformation. the NDI-causing mutant receptors might disrupt the single disulfide bond of the receptor by forming an alternative bridge with either one of the conserved cysteine residues. Because high resolution structural data are not available for the V 2 receptor, we have computed a three-dimensional model with special emphasis on the extracellular domains. Calculations are based on the conformations of identical or similar fragments of other proteins retrieved from the Brookhaven 3D protein data bank (for details of the modeling procedure, see Fig. 1 and "Experimental Procedures"). The conformations of the protein fragments were assembled stepwise on the foundation of our previously described model of the intracellular and transmembrane domains (13), and the stability of the completed model was assessed by molecular dynamics simulations. Top and side views of the structure model are shown in Fig. 2,  A and B, respectively. The positions of the NDI-causing additional cysteine mutations, the conserved cysteine residues Cys-112 and Cys-192 (which form the disulfide bond between ECL1 and ECL2) and the single nonconserved extracellular cysteine residue Cys-195 are indicated. It is thought for the V 2 receptor that AVP binds with its cyclic portion in the large cavity formed by the upper parts of TM3, TM4, TM5 and TM6 and ECL2 and with its C-terminal portion in the smaller cavity formed by TM2, TM3, TM7, and parts of ECL1, ECL2, and ECL3 (20). Similar structure models for the location of the ligand binding domain have been proposed for other GPCRs with peptide ligands (21). Our model predicts that the entrance of the larger ligand binding cavity is surrounded by the long ECL2 domain, which forms a U-like loop (see Fig. 2A). Residue Cys-195 would be located in the center of the extracellular face of the receptor at the tip of the U-like loop and, more importantly, it would be accessible to each of the cysteine residues introduced by the NDI-causing mutations. Our model thus raises the possibility that the additional cysteine residues form a second disulfide bond with residue Cys-195 rather than disrupting the conserved bond between Cys-112 and Cys-192. This second bond may be formed easily in the case of our previously described mutant receptors G185C and R202C (12) because their additional cysteine residues lie at the surface of the molecule at approximately the same level as residue Cys-195 (see Fig. 2B). Formation of the putative second disulfide bond could cause sealing of the entrance of the larger ligand binding cavity and/or may pertubate this site at a deeper level (see Fig. 2A). In either case, the newly formed disulfide bond should not have a major influence on the overall structure of the receptor.

Construction of a Structure
A C195A Mutation Suppresses the Defect of the NDI-causing Mutations G185C and R202C-To address the question of whether or not second disulfide bonds are formed upon introduction of the additional extracellular cysteine residues, we have introduced a C195A mutation into our previously described GFP-tagged NDI-causing receptor mutants G185C.GFP and R202C.GFP (12) (resulting receptor mutants G185C/C195A.GFP and R202C/C195A.GFP). If additional bonds are present, mutation of Cys-195 should disrupt them and lead to functional rescue of the NDI-causing mutant receptors. It was demonstrated previously that mutation of residue Cys-195 alone led to receptors with almost wild-type properties, i.e. that this residue is not necessary for V 2 receptor function (22). 2 The pharmacological properties of the double mutants were characterized in transiently transfected HEK 293 cells by performing [ 3 H]AVP binding assays with intact cells (Fig. 3) and adenylyl cyclase activity assays with crude membrane preparations (Fig. 4). The GFP-tagged wild-type V 2 receptor (WT.GFP) and the NDI-causing single mutants (G185C.GFP and R202C.GFP) were used as controls. As described previously (12), no significant specific [ 3 H]AVP binding was observed for the NDI-causing mutant receptors G185C.GFP and R202C.GFP. The introduction of the C195A mutation, however, led to functional rescue of both mutant receptors. The K d values of both double mutants were comparable to that of WT.GFP (5.2 nM for G185C/C195A.GFP versus 5.4 nM for WT.GFP; 2.1 nM for R202C/C195A.GFP versus 4.7 nM of WT.GFP). Whereas the B max value of G185C/C195A.GFP (0.62 pmol/mg) was also in the range of WT.GFP (0.93 pmol/ mg), that of R202C/C195A.GFP (0.24 pmol/mg) was decreased to 25% of that of WT.GFP (0.98 pmol/mg).
The more sensitive adenylyl cyclase assays revealed AVP-dependent cAMP formation for both NDI-causing single mutants with decreased maximal activities and strongly increased EC 50 values (159 nM for G185C.GFP versus 3.4 nM for WT.GFP; 503 nM for R202C.GFP versus 3.8 nM for WT.GFP). Residual cAMP formation, indicating the presence of low affinity binding sites, was also detected in our previous study for mutant receptor R202C but not for G185C (12). Because untagged receptors were used in the previous study, the adenylyl cyclase-stimulating activity of G185C.GFP is obviously caused by the GFP moiety which may, in this case, cause a stabilization of the mutant receptor. In any case, the adenylyl cyclase activity assays demonstrate the functional rescue of both NDI-causing mutant receptors by additional C195A mutations. The doseresponse curves of both double mutants were left-shifted compared with those of the NDI-causing single mutants (K d values: 3.4 nm for G185C/C195A.GFP versus 159 nM for G185C.GFP; 33, 20 nM for R202C/C195A.GFP versus 503 nM for R202C.GFP). The maximal activities were comparable to that of the wild-type.
The [ 3 H]AVP binding assays with intact cells revealed the B max of the double mutant R202C/C195A.GFP to be significantly lower than that of WT.GFP (see above). This may be the result of decreased expression and/or an impaired cell surface transport caused by the now free cysteine residue at position 202. We therefore analyzed the expression and cellular distribution of the double mutants by localizing their GFP fluorescence signals in transiently transfected HEK 293 cells, using laser scanning microscopy. After recording the GFP fluorescence signals (Fig. 5, left panels in green), the cell surface of the same cells was identified by the use of trypan blue (Fig. 5,  central panels, in red). Trypan blue does not penetrate living cells, and its autofluorescence is suitable for visualizing the cell surface (15). The computer overlay of green GFP fluorescence and red trypan blue fluorescence allows identification of recep-tors that are transported to the cell surface (Fig. 5, right panels, colocalization is indicated by yellow). The GFP signals of the double mutants were compared with that of WT.GFP. The likewise transport-competent single mutant receptors G185C.GFP and R202C.GFP (12) were also used as controls.
GFP signals were detected at the cell surface for WT.GFP and the single mutants G185C.GFP and R202C.GFP, as indicated by the overlap with the trypan blue signal. Additional GFP signals were located inside the cells, presumably representing transport intermediates en route to the cell surface or receptors retained within the cell as a consequence of overexpression. The pictures for the double mutants G185C/ C195A.GFP and R202C/C195A.GFP were very similar to that of WT.GFP, i.e. the receptors seem to be present in comparable amounts at the cell surface. For the double mutant R202C/ C195A.GFP, these results indicate that its decreased B max is caused neither by decreased expression nor decreased transport to the plasma membrane.
Expression and transport of the double mutants in transiently transfected HEK 293 cells were also analyzed by immunoblot detection of total receptors (i.e. intracellular and cell surface receptors) in crude membranes (Fig. 6A) and by monitoring plasma membrane-bound receptors with cell surface biotinylation assays (Fig. 6B), respectively. Cells were transfected and grown in 60-mm-diameter dishes following an identical protocol prior to the assays.
For the detection of the receptors in crude membranes (Fig.   FIG. 5. Localization of the double mutants G185C/C195A.GFP and R202C/C195A.GFP in transiently transfected HEK 293 cells by confocal laser scanning microscopy. Cells expressing the wild-type GFP-tagged receptor (WT.GFP) and the likewise transportcompetent single mutant receptors G185C.GFP and R202C.GFP (12) were used as a control. Cells were analyzed with horizontal (xy) scans. Receptor GFP fluorescence signals are shown in green (left panels) and cell surface trypan blue signals of the same cells in red (central panels). GFP and trypan blue fluorescence signals were computer-overlaid (right panels; overlap is indicated in yellow). GFP fluorescence is detectable only in the case of cells that were successfully transfected, whereas cell surface trypan blue fluorescence is detectable for every cell in the field of view. The scans show representative cells. Scale bar, 25 m. Similar data were obtained in four independent experiments. 6A), a glycosylation state analysis was performed to facilitate identification of intracellular forms. Membranes were treated with endoglycosidase H to remove high mannose glycosylations (indicative of endoplasmic reticulum forms), and PNGase F to remove both high mannose and complex glycosylations (the latter indicative of post endoplasmic reticulum forms). WT.GFP was used as a control for the presence of complex glycosylations. The previously described (13) transport-deficient GFP-tagged V 2 receptor mutant ⌬L62-R64.GFP, which contains a deletion of the sequence 62 LAR 64 in ICL1, was used as a negative control for complex glycosylations. Proteins were detected by immunoblotting using an anti-GFP antiserum. For WT.GFP, two immunoreactive protein bands with apparent molecular masses of 60 -65 kDa and 75-80 kDa were detected in the untreated membranes as described previously (13). The 60 -65-kDa but not the 75-80-kDa bands were sensitive to endoglycosidase H, whereas both bands were sensitive to PN-Gase F. The 60 -65-kDa bands thus represent the high mannose, and the 75-80-kDa bands are the complex glycosylated forms (the complex glycosylated forms do not shift to the nonglycosylated 60-kDa form upon PNGase treatment because O-glycosylations are added to the N terminus of the V 2 receptor in the Golgi apparatus (23)). In contrast to WT.GFP, only the 60 -65-kDa high mannose forms were detectable for mutant ⌬L62-R64.GFP, consistent with its previously described endoplasmic reticulum retention (13). For both double mutants the same pattern and comparable amounts of high mannose and complex glycosylated forms were detected as for WT.GFP. These results show that the decreased B max of R202C/ C195A.GFP did not result from decreased expression, consistent with the laser scanning microscopy localization study.
For the cell surface biotinylation assay (Fig. 6B), plasma membrane proteins of intact cells were labeled with biotin. Biotinylated proteins were isolated with NeutrAvidin and subjected to SDS-polyacrylamide gel electrophoresis. The receptors were detected on immunoblots using an anti-GFP antiserum. Complex glycosylated 75-80-kDa forms were detectable for WT.GFP and both double mutants, but not in the case of the transport-deficient mutant ⌬L62-R64.GFP. For WT.GFP and both double mutants, additional faint 60-kDa protein bands were observed. This protein band seems to represent degradation products of the complex glycosylated forms. It was not observed for mutant ⌬L62-R64.GFP, demonstrating that it does not represent high mannose forms (which might have become labeled e.g. because of cell lysis). The amount of the complex glycosylated forms of double mutant R202C/ C195A.GFP was comparable to that of WT.GFP and the double mutant G185C/C195A.GFP. These results thus show that the decreased B max of R202C/C195A.GFP did not result from a decreased transport level, again consistent with the laser scanning microscopy localization study.
In summary, even though the molecular basis of the decreased B max of R202C/C195A.GFP remains elusive, our results clearly show that the NDI-causing mutant receptors G185C and R202C can be functionally rescued by an additional C195A mutation. They thus strongly suggest that the second disulfide bond predicted by the structure model is indeed formed in the mutant receptors. The Free Extracellular Cysteine Residue Cys-195 of the V 2 Receptor Is Not Conserved throughout Evolution-It was shown previously that residue Cys-195 is not essential for V 2 receptor function (22). Our results demonstrate that this residue might even be disadvantageous if additional cysteine residues are introduced by a mutation. We therefore examined whether residue Cys-195 is conserved throughout evolution. An alignment of the second extracellular loops of the V 2 receptors of various species shows that this residue is not conserved among mammals (Fig. 7A). Most mammals have a basic arginine or histidine residue at this position, which protects them from damage caused by mutationally introduced extracellular cysteine residues. We then determined the approximate time point in evolution when this residue was introduced into the V 2 receptor. We amplified and sequenced the V 2 receptor genes of various primates. Surprisingly, residue Cys-195 is not conserved even among primates (Fig. 7B). New and Old World monkeys have arginine and serine residues at this position, similar to other mammals. Residue Cys-195 seems to have appeared roughly 20 -30 million years ago, because the gibbons, diverging afterward, already possess this residue.
We then asked how many of the class I GPCRs containing a conserved disulfide bond have an uneven number of extracellular cysteines, thus predisposing them to the formation of additional disulfide bonds by mutationally introduced cysteine residues. Among a total of 397 receptors only 18 such receptors (4.5%) were found (Fig. 8), representing only 12 different receptors when species variants are ignored. The presence of an uneven number of extracellular cysteine residues is thus rare in this protein family. DISCUSSION It was demonstrated previously that residues Gly-185 and Arg-202 are not essential for V 2 receptor function (24), indicating that the defects of the NDI-causing mutants G185C and R202C are mediated by the introduced cysteine residues rather than by the residues replaced. The most obvious explanation for these findings was that the additional cysteine residues impair the formation of the conserved disulfide bond between residues Cys-112 and Cys-192. However, we show here by second site C195A mutations that these residues instead seem to participate in a second disulfide bond with residue Cys-195 of the V 2 receptor.
Another hint in the same direction came recently from studies with the neurokinin-1 receptor. Here, introduction of additional cysteine residues into the second extracellular loop did not affect the conserved disulfide bond, and it was suggested that the same may apply to the cysteine-introducing, NDIcausing V 2 receptor mutations (25).
From our functional rescue data it appears obvious that two disulfide bonds are indeed formed in the NDI-causing mutant receptors, although we have not demonstrated directly that this second bond is formed between Cys-195 and an additional cysteine residue. The possibility that two bonds are formed, each involving one of the conserved cysteine residues and either Cys-195 or the mutationally introduced residue (e.g. for the R202C mutant the combinations Cys-112 with Cys-195 and Cys-192 with R202C or Cys-192 with Cys-195 and Cys-112 with R202C) is, at first sight, also consistent with our functional rescue data. But this possibility is very unlikely when our previous results demonstrating that mutation of the conserved cysteine residues leads to a strong transport defect of the mutant receptors are taken into account (12). Each of these alternative combinations would disrupt the conserved disulfide bond and should also cause a transport defect; however, this was not observed. Formation of a second disulfide bond between the additional cysteine residues and Cys-195 is also supported by our molecular modeling data. This bond would cause sealing of the entrance of the ligand binding cavity without having a strong influence on overall receptor folding, consistent with the binding-defective but transport-competent properties of these two mutant receptors (12). The direct demonstration of the second disulfide bonds is nevertheless an important future goal and may be achieved e.g. by receptor isolation, proteolytic digestion, and assignment of disulfide bond-connected fragments by mass spectrometry.
For the NDI-causing mutant receptors G185C and R202C, the additional C195A mutation led to receptors with wild-type K d values for AVP. Whereas the B max of the double mutant G185C/C195A was also in the range of the wild-type, that of mutant R202C/C195A was significantly reduced. The laser scanning microscopy localization studies, glycosylation state analyses, and cell surface biotinylation assays consistently demonstrated that the reduced B max of this double mutant is attributable to neither low expression levels nor decreased transport to the cell surface. A possible explanation for the reduced B max is that the cell surface receptors comprise a mixed population of functional and nonfunctional receptors. The liberated cysteine residue at position 202 may cause an unstable folding state, leading to the formation of two different conformations: one similar to that of the wild-type, the other leading to a binding-defective but transport-competent receptor.
The mutations G185C and R202C are not unique in causing NDI by the introduction of additional cysteine residues. The mutations R106C, R181C, and Y205C belong to the same family. Our structure model also allows predictions of the defects mediated by these mutations. The cysteine residue of mutant receptor R106C would lie at the same level in the molecule as those of the mutant receptors G185C and R202C and may also easily reach residue Cys-195 (see Fig. 2A). The functional properties of mutant receptor R106C have not been described as yet, but formation of a second disulfide bond seems to be very likely in this case. The same may be true for mutant receptor R181C. Although its extra cysteine residue would be located at a deeper level in the molecule, it may also be close enough to Cys-195 to form a second disulfide bond (see Fig. 2B). Consistent with this, impaired ligand binding but preserved transport to the plasma membrane was described for this mutant receptor (5). A binding defect was also described for mutant receptor Y205C (9). In this case, a second disulfide bond may also be formed, although the introduced cysteine residue is located more distantly from Cys-195.
Our data bank analysis demonstrates that the presence of an uneven number of extracellular cysteine residues is rare within FIG. 8. Alignment of the extracellular domains of class I G protein-coupled receptors with an uneven number of extracellular cysteine residues. The data set was constructed from the SWISS-PROT and EMBL data banks. The alignment shows only the sequences of the N-terminal tails (Ntt) and the extracellular loops (ECL1, ECL2, and ECL3). The transitions to the transmembrane domains are indicated (T1-T6). Conserved extracellular cysteines are framed in black, nonconserved cysteine residues in gray. The numbers at the end of the receptor abbreviations (clear frame) indicate the total number of extracellular cysteine residues of the corresponding receptors. Receptor abbreviations were adopted from the original SWISS-PROT or EMBL data files: V2r hum, human vasopressin V 2 receptor; Opsd hum, human rhodopsin; Opsd bov, bovine rhodopsin; Opsd rat, rat rhodopsin; D4dr mou, mouse D4 dopamine receptor; D4dr rat, human D4 dopamine receptor; Rgr bov, bovine RPE retinal G-protein-coupled receptor; Rgr hum, human RPE retinal G-protein-coupled receptor; B1ar mou, mouse ␤1-adrenergic receptor; B1ar rat, rat ␤1-adrenergic receptor; Hh2r hum, human histamine 2 receptor; 5h2a rat, human 5-hydroxytryptamine 2a receptor; rat 5h2b rat, rat 5-hydroxytryptamine 2b receptor; 5h2c hum, human 5-hydroxytryptamine 2c receptor; 5h5b mou5, mouse 5-hydroxytryptamine 5b receptor; 5h5b rat, rat 5-hydroxytryptamine 5b receptor; Ssr1 hum, human somatostatin 1 receptor; Grhr pig, pig gonadotropin-releasing hormone receptor. the class I family of GPCRs containing a conserved disulfide bond. We propose the existence of a strong selection pressure to avoid such a situation because these receptors are predisposed to inactivation by disulfide bond formation following the mutational introduction of extracellular cysteine residues. Interestingly, human rhodopsin was found among the receptors with an uneven number of extracellular cysteines. Here, the Y178C mutation was described as causing retinitis pigmentosa, and, similarly to the corresponding NDI-causing mutations of the V 2 receptor, it was proposed that the Y178C mutation impairs the formation of the conserved disulfide bond of rhodopsin (11). Taking our results into account, the introduced cysteine residue may also form a second disulfide bond with the nonconserved, extracellular cysteine residue of rhodopsin. If so, additional mutation of the nonconserved cysteine residue should also suppress the defect of the Y178C mutation.