Post-translational Modifications of Endothelin Receptor B from Bovine Lungs Analyzed by Mass Spectrometry*

A new mild experimental approach for isolation of peptide membrane receptors and subsequent analysis of post-translational modifications is described. Endothelin receptors A and B were isolated on oligo(dT)-cellulose usingN-(ε-maleimidocaproyloxy)succinimide endothelin coupled to a protected (dA)-30-mer. This allowed a one-step isolation of the receptor from oligo(dT)-cellulose via variation solely of salt concentration. The identity of the receptor was confirmed by direct amino acid sequencing of electroblotted samples or by using antibodies against ETA and ETB receptors. The method used here is very fast, requires only very mild elution conditions and, for the first time, gave both ETA and ETB receptors concurrently in very good yield. Following enzymatic in-gel digestion, MALDI, and electrospray ion trap mass spectrometric analysis of the isolated endothelin B receptor showed phosphorylation at Ser-304, -418, -438, -439, -440, and -441. Further phosphorylation at either Ser-434 or -435 was observed. The endothelin B receptor is also palmitoylated at Cys residues 402 and 404. Phosphorylation of Ser304 may play a role in Hirschsprung’s disease.

A new mild experimental approach for isolation of peptide membrane receptors and subsequent analysis of post-translational modifications is described. Endothelin receptors A and B were isolated on oligo(dT)-cellulose using N-(⑀-maleimidocaproyloxy)succinimide endothelin coupled to a protected (dA)-30-mer. This allowed a one-step isolation of the receptor from oligo(dT)-cellulose via variation solely of salt concentration. The identity of the receptor was confirmed by direct amino acid sequencing of electroblotted samples or by using antibodies against ET A and ET B receptors. The method used here is very fast, requires only very mild elution conditions and, for the first time, gave both ET A and ET B receptors concurrently in very good yield. Following enzymatic in-gel digestion, MALDI, and electrospray ion trap mass spectrometric analysis of the isolated endothelin B receptor showed phosphorylation at Ser-304, -418, -438, -439, -440, and -441. Further phosphorylation at either Ser-434 or -435 was observed. The endothelin B receptor is also palmitoylated at Cys residues 402 and 404. Phosphorylation of Ser 304 may play a role in Hirschsprung's disease.
Although it is clear that post-translational modifications of G-protein-coupled receptors are intimately involved in the physiological function of these signal transduction systems, there is as yet relatively little direct evidence for the specific modifications and the relationship of the different modifications to signal transduction processes. These types of processes are not easily amenable to analysis by genomic methods, i.e. there is a need for efficient methods to directly analyze these processes at the proteome level. We report here new, efficient methods for rapid isolation of the endothelin B receptor and for highly sensitive analysis of its post-translational modifications via mass spectrometry.
Endothelin, the strongest vasoconstrictor yet known, is a 21-amino acid peptide with physiological effects on cellular development, differentiation, vasoconstriction, and mitogenesis (1,2). There are 3 different endothelin isoforms, ET-1, 1 ET-2, and ET-3 with different affinity for the two different endothelin receptor subtypes, A and B (3)(4)(5)(6). Both receptors are members of the G-protein-coupled receptor superfamily (3)(4)(5)(6)7). The consensus ET receptor topology includes three extracellular domains, three intracellular loops, and a cytoplasmic COOH-terminal tail, separated by seven hydrophobic helical regions thought to span the lipid bilayer. In addition it has been presumed that ET receptors are post-translationally modified by glycosylation of the NH 2 terminus and by phosphorylation and palmitoylation of the cytoplasmic surface. Based on homology with other G-protein-coupled receptors (8 -10), there has been speculation regarding possible structures, functional regions, and sites of post-translational modifications for endothelin receptors (11)(12)(13)(14)(15)(16). However, as yet there is very little direct evidence for attributes such as the sites and the roles of glycosylation, palmitoylation, and phosphorylation, the location of the endothelin-binding site and the basis for discrimination among the three different endothelin isoforms.
A role of endothelin in disease has recently been demonstrated by the finding that mutated ET B receptor is associated with Hirschsprung's disease (17)(18)(19)(20)(21). In addition, mice lacking the ET-1 gene display severe malformation of large blood vessels, stressing the importance of endothelin during development (22). Endothelin has been shown to be a mitogenic agonist in different cell types (23,24). The signaling pathway by which ET-1 promotes cell proliferation involves activation of intracellular kinase cascades and transcription factor stimulation. Recently it was suggested that the cytoplasmic tail of ET B receptor is involved in activation of three distinct mitogen-activated signal transduction pathways requiring extracellular-regulated kinase, c-Jun N-terminal kinase, and p38 kinases (25). Studies conducted by sitedirected mutagenesis of ET A receptor suggested that the third intracellular loop and the COOH-terminal tail are also important for receptor-G-protein coupling (26,27).
We report here a new method for isolation of endothelin receptor using oligo(dA) covalently linked to endothelin via a specially developed bifunctional cross-linker. Affinity chromatography has been carried out using oligo(dT) columns with mild elution using only changes in salt concentration analogous to methods used for isolation of eukaryotic mRNA. In-gel digestion of electrophoretically purified receptor, subsequent peptide mass fingerprinting by MALDI-TOF or electrospray ion trap mass spectrometry, and fragment analysis by tandem (MS/MS) mass spectrometry have been used to characterize post-translational modifications of this receptor.

Membrane Preparations
Bovine lung (1.4 kg) was homogenized in a Heavy Duty Blender in 3 volumes of 20 mM Tris-HCl, pH 7.4, which contained 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml pepstatin, and 10 g/ml leupeptin (buffer A). The homogenized lungs were centrifuged at 5000 ϫ g for 20 min at 4°C. The supernatants were discarded, the membrane pellets were suspended in 3 volumes of buffer A and washed 3 times by centrifugation. Afterward the membranes were washed twice in 20 mM phosphate buffer, pH 7.4, which contained 500 mM NaCl, 20 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, and the same amount of protease inhibitors as in buffer A (buffer B). The membrane preparations were stored at Ϫ70°C.

Ligand Binding Assays
The binding activity of the solubilized endothelin receptor was tested with 125 I-endothelin-1. 10 -20 g of total protein were incubated for 2 h at 20°C with 125 I-endothelin-1 (15,000 cpm) in a total volume of 100 l of NaCl/P i , 20 mM sodium phosphate, 150 mM sodium chloride, pH 7.4, containing 1 mM EDTA and 2 mM phenylmethylsulfonyl fluoride. Binding was terminated by gel filtration through Whatman GF/B glass filter sheets precoated with 0.3% (w/v) polyethyleneimine. The filters were washed three times with 4 ml of NaCl/P i buffer in a filtration device (Hoeffer) and transferred to polystyrene tubes. The filter-bound radioactivity was measured in a ␥-counter (Packard).
Derivatized endothelin-1 (EMC-ET) was purified on a Sephacel C18 column using a gradient of 5-70% acetonitrile in 0.1% trifluoroacetic acid. The purity of the EMC-ET containing fraction was checked by protein sequencing and mass spectrometry (found MH ϩ 2685.0, expected 2685.0). The susceptibility of the maleimido group to Michael addition of thiols was indirectly monitored with cysteine and Ellman's reagent (33). Purified EMC-ET was stored at Ϫ20°C.
Synthesis of (dA)30 -5Ј-S-EMC-ET-To (dA)30 -5Ј-SS-R 1-O-dimethoxytrityl hexyldisulfide in 45 l of 100 mM triethylammonium acetate, 10 l of freshly prepared 1 M dithiothreitol in 200 mM Tris/HCl, pH 8.0, were added. The solution was mixed briefly then incubated for 3 h at room temperature to achieve quantitative reduction of the disulfide to the free thiol. For complete removal of the dithiothreitol, a gel filtration on PC 3.2/10 column was performed under isocratic conditions (50 mM MES, pH 6.0, 5 mM EDTA). The fraction with the deprotected oligonucleotide (about 200 l) was immediately mixed with 15 nmol of EMC-ET, which had previously been concentrated in a Speed-vac, and adjusted to pH 6.0 in a volume of 150 l with 200 mM MES. The reaction mixture was incubated overnight under argon to prevent dimerization of the thiol-containing oligonucleotides. The extent of the reaction was

Purification of ET Receptor on Oligo(dT)-cellulose using
Oligo(dA)-coupled Endothelin-1 Frozen membranes (100 g) were thawed and suspended in 2 volumes of 20 mM potassium phosphate, pH 7.4, containing 0.40% digitonin, 0.25% Chaps, 500 mM NaCl, 20 mM EDTA, 1 g/ml RNase, and the same protease inhibitors as buffer A (buffer C). The mixture was gently stirred at 4°C for 2 h, filtered through 3 layers of miracloth and the filtrate was centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatants (85 ml) were incubated with 0.5 nmol of (dA)30 -5Ј-S-EMC-ET for 3 h at room temperature, 100 mg of oligo(dT)-cellulose was added and the suspension was gently agitated at 4°C for 12 h. Oligo(dT)-cellulose was pelleted by centrifugation (4°C, 1,000 ϫ g, 5 min) and packed in a micro column. The column was washed with 10 ml of buffer C at 4°C and afterward eluted with 10 mM Tris/HCl, pH 7.4, 0.40% digitonin, 0.25% Chaps, 1 mM EDTA, 10 g/ml leupeptin, 10 g/ml pepstatin. 200-l fractions were collected. Fractions containing endothelin receptor were identified by SDS-PAGE and by immunoblot analysis.

Immunoblot Analysis
Immunoblot analysis followed the procedure described by Hagiwara et al. (35). 5 l of the protein sample were loaded per gel lane. Proteins from SDS-PAGE were transferred electrophoretically to a nitrocellulose membrane (Scleicher & Schuell). The blot was incubated with anti-ET B receptor serum at a 1:5,000 dilution in 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20, at room temperature for 1 h. The receptorantibody complexes were treated with alkaline phosphatase-conjugate goat anti-rabbit immunoglobulin G antiserum (Sigma) at 1:2.000 dilution. The immunoprecipitate was stained with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (ready made solution from Böhringer, Mannheim).

Mass Spectrometric Analysis
For MALDI mass spectrometry, samples were dissolved in 5 l of 50% acetonitrile, 0.1% trifluoroacetic acid and sonicated for few minutes. Aliquots of 0.5 l were applied onto a target disk and allowed to air dry. Subsequently, 0.3 l of matrix solution (1% w/v ␣-cyano-4hydroxysuccinnamic acid in 50% acetonitrile, 0.1% (v/v) trifluoroacetic acid) was applied to the dried sample and again allowed to dry. Spectra were obtained using a Bruker Biflex MALDI time-of-flight mass spectrometer. MS/MS analyis was carried out using Finnigan Mat (San Jose, CA) LCQ ion trap mass spectrometer. For the interpretation of MS and MS/MS spectra of protein digests we used the Sherpa software (37) and the MS-Fit program available at the www site at the University of California at San Francisco (http://rafael.ucsf.edu/cgi-bin/msfit).

Protein Sequencing of the ET B Receptor
Proteins were electroblotted onto poly(vinylidenedifluoride) membrane, stained with Coomassie Blue, and directly used for protein sequencing. The protein sequence analyses were performed on an Applied Biosystems Procise 494 4-cartridge sequencer (ABI, Foster City). The sequencing results were analyzed using a Macintosh based 610A data system.

Isolation of ET B Receptor Using a New "
Fishhook"-We have designed a heterobifunctional cross-linking reagent which in addition to a N-hydroxysuccinimide-activated ester also possesses a thiol-selective maleimido group. This linker was first attached to endothelin-1 via the ⑀-amino group of Lys 9 . The identity of the N-(⑀-maleimidocaproyloxy)succinimide endothelin (Fig. 1A) was checked by MALDI mass spectrometry and protein sequencing. Following HPLC purification, the thiol selective maleimido group was used to attach 30-mer(dA). After purification (Fig. 1B), the final product, (dA)30 -5Ј-S-EMC-ET, was subsequently added to a suspension of bovine lung membranes solubilized in digitonin/Chaps. The ET receptors were purified in a manner similar to that used for isolation of eukaryotic mRNA: absorbtion on oligo(dT)-cellulose, washing under conditions which favor formation of a (dA⅐dT) double helix (high salt concentration) and elution with very low salt concentrations that destabilize a (dA⅐dT) double helix.
Endothelin A and B receptors were isolated in virtually pure form ( Fig. 2A) and their identity was confirmed by immunoblotting with antibodies raised against ET A and ET B receptors. We estimate that about 30 pmol of ET B receptor was obtained from 100 g of bovine lung tissue. This is more than twice the yields that we and others have previously reported (38). As expected from our previous experience (39,40), ET B receptor was isolated in larger quantities than ET A (Fig. 2A). The specificity of the isolation procedure was further verified by experiments demonstrating that natural endothelin competes with the (dA)30 -5Ј-S-EMC-ET fishhook (Fig. 2B) and that the yield of the receptor depended on the concentration of (dA)30 -5Ј-S-EMC-ET (Fig. 2C). As previously noted by ourselves and others (38,39), two ET B receptor species of 49 and 34 kDa were observed. After electroblotting, sequence analysis gave the NH 2 -terminal sequence EEREFP for the ET B receptor of 49 kDa, which also conforms to previous identification of the NH 2 terminus of the ET B receptor (38,39).
MALDI-TOF Mass Spectrometry-From the SDS-PAGE gels, about 4 pmol of Coomassie Blue-stained ET B receptor were subjected to in-gel tryptic digestion. We took precautions during gel electrophoresis to avoid formation of acrylamide adducts and used only the best and purest chemicals and solvents available throughout the entire purification process. MALDI mass spectrometry was used for initial analysis of the entire mixture of tryptic peptides. This gave predominantly singly charged fragments, which allows easier interpretation of masses observed for peptide mixtures than is the case for spectra generated by electrospray mass spectrometry. From the MALDI mass spectra (Fig. 3), it was possible to identify the entire NH 2 -terminal extracellular region and peptides from the second and third extracellular regions. From the cytoplasmic region, peptides from the second, third, and COOH-terminal loops were observed. Peptides containing transmembrane helical regions joined to loop regions could also be observed (Fig. 4, Table I). Half of the total tryptic digest (about 2 pmol) was submitted to HPLC separation using octyl-␤-glycoside as a detergent during elution. The separated peptides were collected and subjected to MALDI analysis. As summarized in Fig.  4 and Table I HPLC-separated peptides allowed the observation of peptides corresponding to the entire protein sequence including all transmembrane helices. Due to incomplete tryptic cleavage, which is often observed for membrane receptors, two membrane helices with the intervening loop were often recovered in a single peptide, presumably because the attached loops decrease the overall hydrophobicity of the fragment (Fig. 4). The mass measurements were sufficiently accurate (better than 0.1% of the calculated mass) to give clear indications of which peptides were post-translationally modified (Table I). We observed phosphorylation for peptides 31, 45, 48, and 49. Peptides 31, 45, and 48 showed an increased mass of 80 Da, which is characteristic for peptides with single phosphorylation sites. For peptide 49, masses consistent with mono, di-, tri-, and tetra-phosphorylated peptide were observed. For peptide 43, palmitoylation at two sites was suggested by an increase of mass of 476 Da (2 ϫ 238 Da). Although endothelin B receptor contains two consensus glycosylation sites at residues Asn-60 and -118, the masses of peptide 2a-9 (residues 27-100), peptide 4 (50 -65), and peptide 10 -15 (101-164) were consistent with no glycosylation at these sites.
Electrospray Ion-trap Mass Spectrometry-To confirm the identity of the peptides and to determine the attachment sites of the post-translational modifications, we performed electrospray ion trap mass spectrometry of the unseparated peptide mixture with subsequent MS/MS analysis of selected fragments of interest. We observed that in electrospray-ion trap mass spectrometry the phosphorylated peptides partially lose the H 3 PO 4 moiety in the mass spectrometer, thus producing a pair of peaks separated by a mass difference of 98 Da. This has previously been reported for MALDI-ion trap mass spectrometry of phosphorylated peptides (41). In accordance with this, we observed that phosphopeptides could be identified by a pair of masses 80 Da higher (the mass of H 3 PO 4 minus H 2 O) and 18 Da lower than expected based on the amino acid sequence.

DISCUSSION
In the work reported here the special fishhook, (dA)30 -5Ј-S-EMC-ET, allowed for a rapid, very mild single step isolation of ET B receptor by absorbtion to oligo(dT)-cellulose and elution in a small volume of very low salt buffer. The fishhook has been constructed in such a way that it should be applicable to the isolation of other peptide hormone receptors, which we are currently testing. After this single step procedure, the receptor is identified by SDS-gel electrophoresis and the protein from the gel band is subjected to tryptic fragmentation followed by analysis of the peptide mixture by MALDI and electrospray mass spectrometry.
Mass fingerprinting of tryptic peptides is becoming a standard procedure for rapid identification of proteins in proteome analysis (42,43). The essence of this method is that treatment with trypsin produces a limited number of peptides and that the identification of a surprisingly small number of such peptides suffices to identify the protein in sequence data banks without the need for detailed analysis of potentially very complex mass spectroscopic fragmentation patterns. In the present work we have found that mass fingerprinting of tryptic fragments of ET B receptor by MALDI mass spectroscopy also provided an efficient method to screen the entire protein sequence for the presence of post-translational modifications. Peptides covering the entire receptor sequence could be detected (Fig. 4) with adequate mass accuracy (better than 0.1%, Table I) to identify sequential regions containing post-translational modifications. Although it would in principle be possible to use only electrospray ion trap mass spectrometric analysis of the peptide mixture with subsequent MS/MS sequencing of each peptide, the prior use of MALDI mass spectrometry to identify the singly charged peptide masses of interest resulted in a dramatically simplified interpretation of the data. It should be stressed that in the present study, as is the general case with mass spectroscopy of peptides and proteins, a large number of spectra were acquired under different ionization/fragmentation conditions to obtain adequate mass information on all peptides. In this regard, we also found that MALDI analysis of membrane peptides was facilitated by appropriate HPLC of these peptides. Inefficient peptide separation due to aggregation and the loss of very hydrophobic peptides from integral membrane proteins in RP-HPLC is a well known problem. We used octyl-␤-glycoside in separation solvents, which definitively helped in the recovery of transmembrane peptides, but most of the peptides were found in several HPLC fractions. This clearly means that noncovalent aggregation of peptides is still a problem during reverse phase-HPLC. However, this was of no major concern in the present work since MALDI showed that such aggregates dissociated into their peptide components. The presence of octyl-␤-glycoside to keep the membrane peptides soluble resulted in much better signal intensity for these peptides.
Our analysis of peptides from ET B receptor showed that we were able to recover peptides from the entire sequence of the receptor. MALDI, electrospray ion trap mass spectrometry, and MS/MS analysis revealed that ET B receptor is phosphorylated at Ser 304 , Ser 418 , Ser 438 , Ser 439 , Ser 440 , and Ser 441 . Furthermore, another phosphorylation at Ser 434 or Ser 435 was observed. The ET B receptor is also palmitoylated at two sites identified to be at Cys 402 and Cys 404 . A few peptides were observed both with and without phosphorylation. Given the distinctive fragmentation patterns of these post-translational modifications during mass spectroscopy, we believe this reflects the existence of species of ET B receptor with different patterns of post-translational modifications. At the present state of knowledge about G-protein-coupled receptors, the large number of post-translational modifications observed for the ET B receptor was unanticipated and suggests that the number and diversity of post-translational modifications of such receptors, together with concomitant roles in signal transduction pathways, may be much more complex than presently realized.
Phosphorylation and palmitoylation at carboxyl-terminal sites are known to influence the signal transduction in some G-protein-coupled receptors (8 -14). The COOH-terminal tail in all known ET B receptors is identical and with the exception of a few amino acid exchanges, all known ET A receptors also have identical COOH-terminal tails. However, there is almost no homology in the COOH-terminal region between ET A and ET B receptors (3, 4, 7, 44 -46). Among the very few strongly conserved amino acids are Ser 418 and Ser 439 as well as Cys 402 and Cys 404 which we found to be phosphorylated and palmitoylated, respectively. Those positions may be post-translationally modified in the ET A receptor as well. However, phosphorylation of Ser 304 and Tyr 438 is possible only for the ET B receptor since these positions are replaced by Gly or Asp, respectively, in the sequence of ET A receptor. It seems likely that different posttranslational modifications of ET A and ET B receptors will correlate with different signal transduction pathways and we are therefore currently analyzing post-translational modifications of the ET A receptor. We suggest that phosphorylation at Ser 304 of ET B receptor may be particularly important in the development of Hirschsprung's disease (17)(18)(19)(20)(21). Hirschsprung's disease is characterized by the absence of autonomic ganglion cells in the terminal bowel. It is the most common cause of congenital obstruction with an incidence of 1 in 5000 live births. It has been shown that the endothelin-induced signaling pathway is crucial for the development of enteric ganglia and that this pathway is genetically compromised in Hirschsprung's disease. The analysis of the ET B receptor gene in patients with Hirschsprung's disease demonstrated two mutations which resulted in stop codons producing truncated and nonfunctional ET B receptor (20) and a single site mutant that replaced Ser 304 , which is phosphorylated (this work) and strongly conserved in all known ET B receptors, with Asn (47). This suggests that phosphorylation of Ser 304 may play a particularly important role in receptor-mediated signal transduction.
The site-directed mutagenesis of ␣ 2 -and ␤ 2 -adrenoreceptors and endothelin receptor A as well as mass spectrometric investigations of rhodopsin showed that those receptors are palmitoylated at Cys residues in the COOH-terminal tails (14,8,27,48). It is believed that the covalently bound palmitic acid residue becomes intercalated in the membrane bilayer. Prevention of palmitoylation of the ␤-adrenergic receptor produced functional uncoupling of the receptor from the adenylyl cyclase pathway, rapid desensitization in response to its ligand, and increased basal phosphorylation (49). For ␣ 2 -adrenergic receptor the signal transduction pathways were not affected, but prevention of palmitoylation caused a decrease of ligand-promoted down-regulation of the receptor. In the case of nonpalmi-toylated bovine rhodopsin, an increase in signal transduction activity was observed (50). Site-directed mutagenesis of ET A receptor (27) has shown that non-palmitoylated ET A receptor shows no change in ligand binding affinity or stimulation of adenylyl cyclase. However, the lack of palmitoylation was reported to affect phosphatidylinositol hydrolysis by phospholipase C activation after stimulation by endothelin-1. Furthermore, it was observed that the mutated ET A , in contrast to wild type ET A , failed to show a ligand-induced transient increase in cytoplasmatic calcium concentration. We have obtained direct evidence that ET B receptor is palmitoylated at Cys 402 and Cys 404 . The cluster of Cys residues, 399 CLCCWC 404 , is highly conserved in endothelin receptors of the B type. Together with the above results on other receptors, this suggests that palmitoylation plays a role in the regulation of the ET B receptor, although no experimental evidence has so far been reported.
The isolation and direct analysis of the chemical structure of membrane receptors has traditionally been a notoriously difficult task. Known difficulties have included low available amounts, limitations in tryptic or other digestion methods, poor peptide separations, very poor recovery of membrane helical fragments, difficulties in obtaining reliable information on post-translational modifications by protein chemical sequencing, etc. As a consequence, the previously available direct protein analyses of membrane receptors were generally limited to proteins that were readily available in larger quantities. Apart from pioneering work on rhodopsin (11,12), there are only a very few scattered reports in the literature on direct observation of post-translational modifications of G-protein-coupled membrane receptors (13)(14)(15). The present results indicate that with the development of new methods for rapid, mild isolation of as little as 1-2 pmol of membrane receptors on gels and the combination of highly sensitive MALDI and electrospray ion trap mass spectrometry, direct evaluation of the sites and types of post-translational modifications of membrane receptors is poised to become a routine characterization. The next challenge will be to correlate different patterns of post-translational modifications with different functional states.