Novel Mutations of the Endothelin B Receptor Gene in Patients with Hirschsprung’s Disease and Their Characterization*

Hirschsprung’s disease (HSCR) is a congenital intestinal disease, characterized by the absence of ganglion cells in the distal portion of the intestinal tract. Recently, three susceptibility genes have been identified in HSCR, namely theRET protooncogene, the endothelin B (ETB) receptor gene (EDNRB), and the endothelin-3 (ET-3) gene (EDN3). To investigate whether mutations inEDNRB could be related with HSCR in non-inbred populations in Japan, we examined alterations of the gene in 31 isolated patients. Three novel mutations were detected as follows: two transversions, A to T and C to A at nucleotides 311 (N104I) and 1170 (S390R), respectively, and a transition, T to C at nucleotide 325 (C109R). To analyze functions of these mutant receptors, they were expressed in Chinese hamster ovary cells. S390R mutation did not change the binding affinities but caused the decreases in the ligand-induced increment of intracellular calcium and in the inhibition of adenylyl cyclase activity, showing the impairment of the intracellular signaling. C109R receptors were proved to be localized near the nuclei as an unusual 44-kDa protein with the extremely low affinity to endothelin-1 (ET-1) and not to be translocated into the plasma membrane. On the other hand, N104I receptors showed almost the same binding affinities and functional properties as those of the wild type. Therefore, we conclude that S390R and C109R mutations could cause HSCR but that N104I mutation might be polymorphous.

disorder, characterized by the absence of ganglion cells in the distal portion of the intestinal tract, as a consequence of premature arrest of cranio-caudal migration of neural crest cells. Recently, three susceptibility genes have been identified in HSCR, namely the RET protooncogene, the endothelin B (ET B ) receptor gene (EDNRB), and the endothelin-3 (ET-3) gene (EDN3).
Mice with targeted null disruption of Ret exhibited an autosomal recessive phenotype comprising the total lack of the enteric nervous system and renal agenesis (1), that provided an excellent model of HSCR. Approximately 50 mutations in RET have so far been identified, and the RET mutations account for 50% of familial and 15-20% of sporadic cases of HSCR, over 75% of which are associated with long segment HSCR (2)(3)(4)(5)(6).
Endothelins (ETs) are a family of 21 amino acid vasoactive peptides, consisting of three isopeptides, ET-1, ET-2 and ET-3, that act on G protein-coupled receptors, ET A and ET B (7)(8)(9)(10)(11). ET A exhibits different affinities to the three isopeptides in the order of ET-1 Ն ET-2 Ͼ Ͼ ET-3. ET B accepts all three isopeptides equally. These receptors regulate multiple effector pathways, for example, phospholipase C via G q , and adenylyl cyclase via G s in smooth muscle cells (ET A ), and via G i in endothelial cells (ET B ) (12,13). Mice with targeted null disruption of either Ednrb or Edn3 exhibit an identical phenotype, coat color spotting, and aganglionic megacolon, similar to HSCR or Shah-Waardenburg syndrome in humans (14,15). Some naturally occurring HSCRs, in humans (a large, inbred Mennonite kindred), mice (piebald-lethal and lethal spotting), or rats (spotting lethal), are related with mutations in EDNRB or EDN3 (14 -18). In isolated patients of HSCR, different mutations of EDNRB or EDN3 have also been reported (19 -24), but no functional analyses of the mutant receptors have so far been performed. To investigate whether mutations in EDNRB are related with HSCR in non-inbred populations in Japan, we examined alterations of the gene in 31 patients and analyzed functional properties of the mutant ET B receptors.

EXPERIMENTAL PROCEDURES
Materials-LipofectAMINE and Trizol were from Life Technologies, Inc. (Tokyo, Japan). ET-1 and ET-3 were from Peptide Institute (Osaka, Japan). 125 I-NaI was from ICN Biomedicals, Inc. (Costa Mesa, CA), and fluorescein isothiocyanate-conjugated goat anti-rabbit F(abЈ) 2 was from ICN Pharmaceuticals, Inc. (Costa Mesa, CA). cAMP radioimmunoassay kit was from Yamasa (Chiba, Japan). Fura-2/AM was from Dojin Chemicals Institute Co. (Kumamoto, Japan). Platelet activating factor, Geneticin, forskolin, (p-amidinophenyl)methanesulfonyl fluoride, peroxidase-conjugated goat anti-rabbit IgG, and 3-isobutyl-1-methylxanthine were from Wako Chemicals (Osaka). Sequenase 7-deaza-dGTP DNA sequencing kit was from U. S. Biochemical Corp. Western blot chemiluminescence reagent was from NEN Life Science Products. * The costs of publication of this article were 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.
Mutagenesis-To construct the cDNAs coding the mutant hET B receptors by cassette mutagenesis (27), we used SR␣ promoter-based mammalian expression vector pME18SfϪ, which carried a cDNA construct encoding the wild-type hET B receptor (28) (kindly provided by Dr. A. Sakamoto, The Institute of Physical and Chemical Research, Saitama, Japan). Mutagenic oligodeoxynucleotide cassettes were created by recombinant PCR with primers (see Fig. 1) (29). The plasmids containing the cDNAs encoding the wild-type hET B receptor and the mutated cassette fragments were treated with the unique cognate restriction enzymes and then ligated (Fig. 1). The mutant plasmids were identified by restriction analysis and verified by automated DNA sequencing.
Transient Expression of the Wild-type and the Mutant hET B Receptors-CHO and Ltk Ϫ cells were maintained in minimum essential medium ␣ and minimum essential medium supplemented with 10% fetal bovine serum, respectively. The cDNAs of the wild-type and the mutant hET B receptors were transfected into CHO or Ltk Ϫ cells in 24-well plates using LipofectAMINE. At 48 -72 h after transfection, the cells were used for the binding assay.
Stable Expression of the Wild-type and the Mutant hET B Receptors-Each expression plasmid was cotransfected with pSV2neo plasmid into CHO cells as described above. Clonal cell lines resistant to Geneticin (0.5 mg/ml) were isolated by colony lifting and screened for the presence of specific 125 I-ET-1 binding.
Radioligand Binding Experiment-125 I-ET-1 was prepared as described before (30). For the competitive binding experiments, the CHO cells transiently expressing each receptor in 24-well plates were incubated at 4°C for 3 h in buffer B (minimum essential medium ␣ containing 10 mM Hepes (pH 7.4) and 0.1% bovine serum albumin) containing 100 pM 125 I-ET-1 with various concentrations of unlabeled ET-1 or ET-3. The cells were then transferred in 0.4 M NaOH to a gamma counter. The density of the receptors was determined by saturation isotherms of 125 I-ET-1 binding to the intact cells as described above, except a condition of incubation at 4°C for 7 h. In membrane binding experiments, to separate bound from free ligands, membrane pellets were washed and centrifuged at 21,800 ϫ g for 5 min three times. Nonspecific binding was defined as the binding in the presence of 100 nM unlabeled ET-1.
Measurement of Intracellular Ca 2ϩ Transient-The cells were incubated in the loading buffer (20 mM Hepes-Hanks (pH 7.4) containing 11 mM glucose and 0.1% bovine serum albumin) with 4 M Fura-2/AM for 60 min at 20°C. Intracellular Ca 2ϩ transients evoked by ET-1 and ET-3 were monitored by a CAF-110 fluorescence spectrophotometer (JASCO Co., Tokyo, Japan) with dual excitation at 340 nm/380 nm and emission at 500 nm.
cAMP Accumulation Assay-The cells at 80% confluency were incubated in 300 l of buffer B containing 1 mM 3-isobutyl-1-methylxanthine at 37°C for 20 min. Agonists and 1 M forskolin were added, and the cAMP accumulation was allowed to continue for 10 min at 37°C. The cAMP contents were measured using a Yamasa radioimmunoassay kit.
Northern Blot Analysis-Total RNA was extracted from the cells using Trizol. The total RNAs (20 g) were electrophoresed on a 1% denaturing agarose gel (containing 18% formaldehyde) and blotted to a nylon membrane. The blots were probed with a 536-bp (EcoRV-EcoRI restriction) nick-translated hET B receptor cDNA fragment and subjected to autoradiography at Ϫ80°C for 34 h. The intensities of specific mRNAs were quantified using a BAS 2000 Fuji Bio-Image analyzer (Fuji Photo Film Co., Kawasaki, Japan).
Preparation of the Anti-peptide Antibody Specific for the hET B Receptor-A peptide (CLKFKANDHGYDNFRSSNKYSSS) corresponding to the carboxyl terminus (420 -442) of the hET B receptor was coupled to keyhole limpet hemocyanin. The antiserum was raised in Japanese White rabbits (2.5 kg). The peptide keyhole limpet hemocyanin was emulsified with Freund's complete adjuvant and injected subcutaneously at multiple sites on the back five times biweekly. The serum was affinity purified with Sulfolink Coupling Gel (Pierce) coupled with the peptide. The antibody recognized only the ET B receptor and not the ET A receptor. The details will be published elsewhere. 2 Immunostaining-The cells were fixed for 1 h in PBS containing 4% paraformaldehyde and then immersed in 50% ethanol in PBS for 10 min. After blocking for 1 h with 10% goat serum in PBS, the cells were incubated with 4 g/ml anti-hET B receptor antibody for 12 h at 4°C. After washing, the cells were incubated in PBS containing 1% goat serum and the fluorescein isothiocyanate-conjugated goat anti-rabbit F(abЈ) 2 fragments for 3 h at room temperature. Fluorescence images were taken in a confocal microscope (Zeiss LSM; Carl Zeiss, Thornwood, NY) using a ϫ 63 objective lens.
Western Blot Analysis and Low Temperature SDS-PAGE-The membrane samples were separated by SDS-PAGE, electroblotted to a polyvinylidene difluoride membrane, and blocked with 4% skim milk for 12 h. After washing, the membrane was incubated with 4 g/ml anti-hET B receptor antibody for 4 h at room temperature and followed by incubation with 0.3 g/ml peroxidase-conjugated goat anti-rabbit IgG for 1 h. The bands were visualized by the Western blot chemiluminescence reagent and quantified using the NIH Image program. The low temperature SDS-PAGE was performed as described (31). Briefly, the membrane protein samples (20 g) were incubated at 4°C for 3 h in buffer B containing 100 pM 125 I-ET-1 with or without 100 nM unlabeled ET-1. After washing, the samples were subjected to SDS-PAGE at 4°C and to autoradiography for 35 h.

RESULTS
SSCP and DNA Sequence Analyses-We analyzed alterations of EDNRB in 31 patients with HSCR (Table I); the ratio of males is 74% (23:31), and those of short type and familial HSCR are 71% (15:21, length undetermined in 10 cases) and 6% (2:31), respectively. In SSCP analyses of the seven exons containing the entire coding region of EDNRB, we identified three mobility shifts, two in exon 1, and one in exon 6 ( Fig. 2A). The allelic status of patients 19, 20, and 16 were heterozygous with normal and aberrant bands, homozygous consisting of aberrant bands and heterozygous with normal and aberrant 2 M. Takimoto and S. Hori, manuscript in preparation. bands, respectively. Nucleotide sequence analysis showed two transversions, A to T at nucleotide 311 and C to A at nucleotide 1170, and a transition, T to C at nucleotide 325 (Fig. 2B). All mutations were missense; the first transversion resulted in the substitution of isoleucine for asparagine at position 104 (N104I), the second, arginine for serine at position 390 (S390R), and the transition, arginine for cysteine at position 109 (C109R). N104I and C109R mutations are located within the transmembrane domain (TM) 1, and S390R is located in the carboxyl-terminal region and adjacent to TM 7.
Ligand Affinity of the Wild-type or the Mutant hET B Receptors-To clarify whether these mutations were related with HSCR predisposition, we examined binding affinities for ET-1 and ET-3 by the competitive binding experiments using intact CHO cells transiently expressing the mutant hET B receptors. The IC 50 values of the wild-type, N104I, and S390R receptors were as follows: 147 Ϯ 39, 240 Ϯ 81, and 159 Ϯ 22 pM for ET-1; 492 Ϯ 120, 339 Ϯ 143, and 328 Ϯ 135 pM for ET-3, respectively (Table II). Similar IC 50 values were obtained in the stably expressing cells (data not shown). These results indicated that N104I and S390R receptors have the ligand affinities closely similar to those of the wild type. In contrast, CHO or Ltk Ϫ cells transfected with C109R receptors showed almost no specific 125 I-ET-1 binding at 100 pM.
Ligand-induced Intracellular Ca 2ϩ Transient and Inhibition of Adenylyl Cyclase Activity via the Wild-type or the Mutant hET B Receptors-To analyze functional properties of the mutant receptors, we used CHO clonal cell lines stably expressing the wild-type (CHO W ), N104I (CHO N104I ), and S390R (CHO S390R ) receptors whose densities were 8.6, 3.8, and 33.3 fmol/10 5 cells, respectively (Table II). In CHO W and CHO N104I , ET-1 elicited similar concentration-dependent increases in [Ca 2ϩ ] i transients (Fig. 3A, Table II). In contrast, the [Ca 2ϩ ] i responses in CHO S390R were markedly decreased, despite the fact that the receptor density was higher than that of CHO W . The EC 50 values of CHO W and CHO S390R were 146 Ϯ 74 and 4,639 Ϯ 1,468 pM, and their maximal responses (E max ) were 157.6 Ϯ 4.5 and 15.8 Ϯ 2.3%, respectively.
In CHO W , ET-1 or ET-3 inhibited forskolin-stimulated adenylyl cyclase (Fig. 3B, Table II). The EC 50 value for the inhibitory effect of ET-1 was 10 pM, and the E max values for ET-1 and ET-3 were 42.6 and 53.5% inhibition, respectively. In CHO N104I , the EC 50 and the E max values for ET-1 and ET-3 were similar to those in CHO W . On the other hand, in CHO S390R , the inhibitory effects of both ET-1 and ET-3 were markedly reduced. Namely the EC 50 for ET-1 was 65 pM, and the E max values for ET-1 and ET-3 were 9.6 and 0%, respectively.
Properties of C109R Receptors-During the course of screening for colonies stably expressing C109R receptors, none of the colonies showed specific 125 I-ET-1 binding at 100 pM, as observed in the cells transiently expressing C109R receptors. To test a possibility that this phenomenon was due to the low levels of the mRNA, a CHO cell line stably expressing C109R receptors (CHO C109R ) was selected, and the amounts of the mRNA and the receptor proteins in the plasma membrane in CHO C109R were examined by Northern blotting and binding analyses. The mRNA level in CHO C109R was proven to be about seven times higher than that in CHO W . In the intact CHO W and CHO C109R cells, the B max values were 3.48 and 0.15 fmol/ 10 5 cells, and the K d values were 0.17 and 0.96 nM, respectively (Fig. 4). On the other hand, in CHO C109R membrane prepara-

FIG. 2. SSCP analysis of exons 1 and 6 of EDNRB in patients with HSCR and sequence analysis.
A, denatured PCR products were electrophoresed at 4°C using a 25 mM Tris, 192 mM glycine buffer, followed by silver staining . N was a normal control. a, b, #19, #20, and #16 were patients with HSCR. B, sequence analysis of #19, #20, and #16.
tions, the displacement analysis of 125 I-ET-1 bindings at 1 nM by unlabeled ET-1 revealed the surprisingly low affinity (the IC 50 value was 39.5 Ϯ 19.8 nM), in no agreement with the case of the intact cells.
Immunostaining with the Anti-hET B Receptor Antibody of the Wild-type and the C109R Cells-To obtain further evidence that the mRNAs were actually translated, the CHO cells were subjected to immunostaining with the anti-hET B receptor antibody. In CHO W , the translated hET B receptors existed diffusely in the entire cells, especially densely in cell processes (Fig. 5A). On the other hand, in CHO C109R , only the regions near the nuclei were stained (Fig. 5B). These stainings were not observed in mock-transfected CHO cells (Fig. 5C). Western blot analysis showed two major bands of 50 and 33 kDa, and a minor band of 44 kDa in CHO W . In contrast, in CHO C109R , one major band of 44 kDa and a faint band of 50 kDa were detected (Fig. 5D). The relative amounts of the 44-kDa protein were 15.3% in CHO W and 80.1% in CHO C109R . Therefore, the 44-kDa protein might account for the staining nearby the nuclei in CHO C109R . These bands were all absent in mock-transfected cells.
Since the ligand-ET B receptor complex is known to be SDSresistant and stable at a reduced temperature, the low temperature SDS-PAGE is a useful method for detection of the active ET B receptor species (31). In the low temperature SDS-PAGE of CHO W membranes, only two bands of 50 and 33 kDa, but not 44 kDa, were detected with 125 I-ET-1 (Fig. 5E), suggesting that the 44-kDa protein was unable to bind to 125 I-ET-1 at 100 pM. DISCUSSION In this study, we analyzed alterations of EDNRB in 31 patients with HSCR and detected three novel mutations: two transversions, A to T and C to A, at nucleotide 311 (N104I) and 1170 (S390R), respectively, and a transition, T to C, at nucle-

FIG. 4. Northern blot analysis (A), and Scatchard analysis of the CHO cells stably expressing the wild-type (B) or C109R (C) ET B receptors.
A, the total RNAs (20 g) were electrophoresed, blotted, and hybridized as described under "Experimental Procedures." A 32 P-labeled 536-bp restriction fragment of the hET B receptor cDNA was used to detect the mRNA levels specific to the hET B receptor. Lane 1, mock-transfected CHO cells; lane 2, the CHO cells expressing C109R receptors; lane 3, the CHO cells expressing the wild type. B and C, the CHO cells expressing the wild-type (B) or C109R (C) hET B receptors, which were used for Northern blot analysis. The cells were incubated with various concentrations of 125 I-ET-1 (31-1,000 pM for the wild type, 42-2,700 pM for C109R). otide 325 (C109R). ET B is a heptahelical receptor that equally binds all three endothelin isopeptides and is involved in the intracellular signaling pathway via heterotrimeric G proteins (17,18). The analyses of both intracellular Ca 2ϩ transients and adenylyl cyclase activities of these mutants revealed that S390R mutation significantly decreased the responses to ET-1, although the binding affinities to the ligands were as high as the wild type. C109R receptors exhibited almost no binding to ET-1 and no functional activities due to the lack of the functionally active receptors in the plasma membrane, because the mutant receptors localized mainly around the nuclei. On the other hands, N104I receptors showed almost the same characteristics as those of the wild type. Taken together, it is concluded that S390R mutation causes a disorder of the intracellular signaling after the binding with agonists, that C109R mutation impairs the translocation of the receptor proteins to the plasma membrane, and that N104I mutation seems polymorphous.
The newly found mutations of EDNRB in this report account for about 6% (2:31) of HSCR, consistent with the previous report (11). Their inheritance states and phenotypes were as follows: one was heterozygous and long segment HSCR, and the other was homozygous and short segment HSCR. Both patients had no associated disease. As described in previous reports (19,(25)(26)(27), these situations are in agreement with a hypothesis that HSCR would be a multigenic disorder. Namely an EDNRB mutation is neither necessary nor sufficient for explaining all clinical HSCR cases but, at least, it possibly accounts for some of them. It is well established that the disruption of the genes encoding the Ret or ET B receptor causes HSCR. Consequently, it is reasonable to expect that mutations of the genes of molecules essential for the signaling pathways via these receptors, such as GDNF, GDNF receptor-␣ (32), neurturin, neurturin receptor-␣ (33, 34), ET-3, and ECE, might lead to HSCR. In fact, mutations in EDN3 caused HSCR (28,29). However, a mutation in GDNF itself was neither necessary nor sufficient to cause HSCR (35,36), because major defects in the Ret pathway would be embryonic lethal or roles of GDNF in enteric neurogenesis could be compensated for by other neurotrophic factors, for example, neurturin. The analysis of another HSCR model mouse, the Dominant megacolon (Dom), might introduce a novel susceptibility gene that lies in human chromosome 22q12-q13 (37).
In this report, we performed functional analyses of the mutant hET B receptors. S390R mutation prevented the mutant receptors from coupling with heterotrimeric G proteins, resulting in decreases in the functions. The amino-terminal segment of the cytoplasmic domain from amino acid residues 390 -402 contains the putative palmitoylation site and forms a fourth cytoplasmic loop, as found in other G protein-coupled receptors including the ET A receptor (38). This loop in the ET A receptor is required to couple with G q but not G s . In coupling of the ET A receptor with G q , this requirement is also revealed (39). The S390R mutation found in the present study, the replacement of serine 390 with arginine, a positively charged amino acid residue, might cause conformational change of this fourth cytoplasmic loop and might decrease in coupling with G proteins.
C109R mutation that occurred in the amino-terminal region of TM 1 impaired the translocation of the mutant hET B receptor into the plasma membrane. Why C109R receptors are not transported into the plasma membrane with their proper orientation? It is generally accepted that when integral membrane proteins are inserted into the endoplasmic reticulum membrane, about 20 amino acids of the amino-terminal region would play a role as a signal sequence and lead the translated polypeptide into the luminal space of the endoplasmic reticulum. It is also known that TM 1 would play a topogenic role as a stop transfer sequence (40,41) and that these segments would require sufficient hydrophobicity to act efficiently. For example, when the non-polar amino acid residue in the type 1 signal anchor sequence of cytochrome P450 was replaced with the basic amino acid residue, the signal anchor sequence was converted to the secretory signal sequence (42). In the case of C109R mutation in this study, the replacement of cysteine with arginine in TM 1 might change the targeting function of TM 1 both by the positively charged group and by the shortened length of the hydrophobic stretch of TM 1. Thus it seems possible that this mutant receptor might deviate from the normal transport pathway from the endoplasmic reticulum to the plasma membrane because of the improper folding and/or membrane topology, resulting in the accumulation near the nuclei (Fig. 5B). A similar example of the hamster ␤-adrenergic receptor was reported previously (43).
It is well known that the normal processing of the wild-type receptors generates a 50-kDa protein and its further degraded form, a 33-kDa protein, by metal proteinases (44). In addition to the wrong translocation as described above, Western blot analysis showed that C109R receptor was processed to an unusual 44-kDa protein and a 50-kDa protein, as a major and a FIG. 5. Immunofluorescence analysis, Western blot analysis, and low temperature SDS-PAGE of the CHO cells stably expressing the wild-type or C109R hET B receptors. The cells were fixed with formaldehyde and permeabilized with 50% ethanol. The hET B receptors were labeled with the anti-hET B receptor antibody (4 g/ml) and then labeled with fluorescein isothiocyanate-conjugated secondary anti-rabbit F(abЈ) 2 fragment antibody. A, the wild-type; B, C109R; C, mock-transfected CHO cells. The scale bars indicate 100 m. D, the membrane protein samples (20 g) were electrophoresed on a 10% SDS-polyacrylamide gel, blotted, and incubated with the antibody (4 g/ml, for 4 h at room temperature) and developed with Western blot chemiluminescence reagent. Lane 1, mock-transfected CHO cells; lane 2, the CHO cells expressing C109R receptors; lane 3, another CHO cells expressing the wild type. E, membranes (20 g) prepared from the CHO cells expressing the wild-type hET B receptor, which was used for Western blot analysis, was incubated at 4°C for 3 h in buffer B containing 100 pM 125 I-ET-1 with (lane 2) or without (lane 1) 100 nM unlabeled ET-1 and washed five times. The samples were subjected to SDS-PAGE at 4°C and to autoradiography for 35 h. Positions of molecular mass standards (in kDa) are indicated. minor products, respectively. In the present study, this 44-kDa protein was also observed in the wild type, although it was a minor component. The binding studies revealed C109R receptors to have the very low B max and the fairly high affinity in the surface of intact cells but the surprisingly low affinity in the total membrane preparations, which include plasma membranes and subcellular organellar membranes (Fig. 4C). Together with the relative ratios of the 44-and 50-kDa proteins in CHO W and CHO C109R cells, these results strongly indicate that the 50-kDa protein, which is located in the plasma membrane as a major form in CHO W and as a minor component in CHO C109R , has the high affinity and that the 44-kDa protein, which is located near the nuclei as a major component in CHO C109R and as a minor component in CHO W , has the very low affinity.
In this study, we identified two mutations in EDNRB susceptible to HSCR and demonstrated that the functional defects of the mutant receptors could be due to not only the impairment of signaling pathways but also the unusual processing pathway and the wrong translocation.