Proc. Natl. Acad. Sci. USA Vol. 89, pp. 7801-7805, August 1992 Biochemistry Direct identification of residues of the receptor in close proximity to the amino terminus of bound epidermal growth factor (cross-linking/ sequendng/binding site) RANDALL L. WOLTJER*t, THOMAS J. LuKASt, AND JAMES V. STAROS*§¶ Departments of *Biochemistry, tPharmacology, and Molecular Biology, Vanderbilt University, Nashville, TN 37235 Communicated by Stanley Cohen, May 19, 1992

ABSTRACT We have recently developed a kinetically con- It is convenient to define the extracellular portion of the trolled, step-wise affinity cross-linking technique for specific, human EGF receptor in terms of the following domains (26): high-yield, covalent linkage ofmurine epidermal growth factor a cysteine-poor N-terminal region (domain I, residues 1-146), (mEGF) via its N terminus to the EGF receptor. EGF receptor and a second cysteine-poor region (domain III, residues from A431 cells was cross-linked to radiolabeled mEGF (MI- 333-460) that separates two cysteine-rich regions (domain II, mEGF) by this technique and the 12'I-mEGF-receptor complex residues 147-332; domain IV, residues 461-621). The bulk of was purified and denatured. Tryptic digestion of this prepa- existing evidence, including the results of all cross-linking ration gave rise to a unique radiolabeled peptide that did not studies reported to date, points to domain III as containing comigrate with trypsin-treated 125I-mEGF in SDS/Tricine gels determinants for receptor binding to EGF (20, 23, 24, 26, 27); but that could be immunoprecipitated with antibodies to however, some studies have also implicated domain I (23, mEGF. The immunoprecipitated peptide was isolated by elec- 25). trophoresis in SDS/Tricine gels, eluted, and sequenced. The We have recently developed a kinetically controlled, step- sequence was found to correspond to that of a tryptic peptide wise affinity cross-linking technique to achieve high yields of ofthe EGF receptor bginning with Gly-85, which is in domain the EGF receptor covalently linked to the N terminus of I, a region N terminal to the first cysteine-rich region of the murine EGF (mEGF) (30). Here we report purification of receptor. Selective loss of signal in the 17th sequencing cycle cross-linked species and conditions for denaturation and suggests that the point of attachment of N-terminally modified tryptic digestion of 125I-mEGF-linked receptor. A radiola- 125I-mEGF to the receptor is Tyr-101 The data presented here beled receptor fragment that did not comigrate in SDS/ provide identification by direct protein microsequencing of a Tricine gels with products resulting from tryptic digestion of site of interaction of EGF and the EGF receptor. 125I-mEGF was further purified; direct protein microsequenc- ing revealed the uniquely migrating peptide to be derived Epidermal growth factor (EGF), a 6040-Da, single-chain from domain I of the extracellular portion of the EGF polypeptide hormone (1, 2), binds to specific membrane receptor. 11 receptors in target cells to exert effects on cell growth and differentiation (reviewed in ref. 3). Rapid effects of EGF MATERIALS AND METHODS binding to its receptor, a single-chain (170-kDa) transmem- brane glycoprotein, include receptor dimerization (4-9) and Materials. mEGF was prepared as described (31); 1251- stimulation ofaprotein tyrosine kinase (10, 11) intrinsic to the mEGF was prepared as described (32), using 1 mg of unla- EGF receptor (12-14), which gives rise to receptor autophos- beled mEGF per ml as a carrier for radiolabeled ligand. The phorylation (12, 15) and phosphorylation of intracellular cross-linking reagent sulfo-N-succinimidyl-4-(fluorosulfon- substrates (reviewed in ref. 3). Much attention has been yl)benzoate (SSFSB) was synthesized as described (30). focused on the interaction of EGF with particular receptor A431 cells were grown to confluence in Dulbecco's modified sites and the relationship of this interaction to short- and Eagle's medium (GIBCO) supplemented with 10%o calf serum long-term biological responses (reviewed in refs. 3 and 16- (GIBCO). Trypsin/EDTA (lx) was from GIBCO. Shed 18). membrane vesicles from A431 cells were prepared as de- A variety of techniques have been used in past work to scribed (15). ATP was purchased from Boehringer Mann- investigate the ligand-binding region of the EGF receptor, heim. Triton X-100 was obtained from Aldrich, and glycerol including preparation of antibodies against the receptor that was from Fisher. Anti-phosphotyrosyl antibody (APY) was compete with EGF for binding (19-23), and functional anal- prepared as described (33, 34) and was used as purified ysis ofchicken/human receptor chimera (23, 24) and receptor antibody coupled to cyanogen bromide-activated Sepharose deletion mutants (25). Covalent cross-linking of 1251-labeled 4B (35). Agarose-bound wheat germ lectin (WGL) was ob- EGF (125I-EGF) to the receptor has been used in previous tained from Vector Laboratories, and NN',N"-triacetylchi- work, in which identification ofthe portion ofthe receptor to which 125I-EGF was linked was deduced from the electro- Abbreviations: APY, Sepharose-coupled anti-phosphotyrosyl anti- phoretic mobility and immunochemical reactivity of the body; DSS, disuccinimidyl suberate; EGF, epidermal growth factor; labeled mEGF, murine EGF; PAS, protein A-Sepharose CL-4B; PTH, products of proteolytic and glycosylytic digests (26, phenylthiohydantoin; SSFSB, sulfo-N-succinimidyl-4-(fluorosulfo- 27). Most recently, EGF bound to the soluble, extracyto- nyl)benzoate; WGL, agarose-bound wheat germ lectin. plasmic portion of the receptor has been visualized in elec- tPresent address: Department of Molecular Biology, Vanderbilt tron microscopic images (28) and crystallized for x-ray dif- University, Nashville, TN 37235. fraction studies (29). ITo whom reprint requests should be addressed at: Department of Molecular Biology, Vanderbilt University, Box 1820, Station B, Nashville, TN 37235. The publication costs of this article were defrayed in part by page charge IA preliminary account of some of this work was presented at the payment. This article must therefore be hereby marked "advertisement" 1990 Annual Meeting ofthe American Society for Biochemistry and in accordance with 18 U.S.C. §1734 solely to indicate this fact. Molecular Biology (47). 7801 Downloaded by guest on September 28, 2021 7802 Biochemistry: Woltjer et al. Proc. Natl. Acad. Sci. USA 89 (1992) totriose was from Sigma. lodoacetamide and dithiothreitol WGL Affinity Purification of Eluate from the Anti- were purchased from Fluka; trypsin (L-1-tosylamido-2- Phosphotyrosyl Resin. A total of 16 ml of pooled APY eluate phenylethyl chloromethyl ketone treated) was obtained from was applied to 0.5 ml of hydrated WGL resin, concentrated Worthington. Anti-EGF immune antiserum was a generous MgCl2 was added to a final concentration of 10 mM, and the gift from S. Cohen (Vanderbilt University). Protein A-Seph- slurry was rocked overnight at 4TC. The WGL resin was arose CL-4B (PAS) was from Pharmacia. Forgels from which sedimented by using the same centrifugation techniques as peptides were isolated for sequencing, acrylamide and NN'- with APY resin described above, and the supernatant was methylenebisacrylamide from BDH were used; for other discarded. Resin-bound receptor was washed by resuspen- gels, these were obtained from Bio-Rad. SDS was purchased sion in 4 ml of ice-cold WGL wash buffer containing 20 mM from Serva, and ammonium persulfate and N,N,N',N'- Hepes (pH 7.4), 10%1 glycerol, 0.2 M NaCl, and 10 mM MgCl2 tetramethylethylenediamine were from Bio-Rad. Mercap- and pelleted as described above; the supernatant was dis- toacetic acid (sodium salt) was obtained from Aldrich. All carded. Elution was accomplished by rocking the resin for 6 other chemicals were obtained from Sigma and were reagent h in 0.5 ml of WGL wash buffer without MgCl2 but supple- grade or better. mented with 3 mM NN',N"-triacetylchitotriose, followed by Affinity Cross-Linking of 125I-mEGF to the EGF Receptor in centrifugation as described above for the resin, and collection A431 Cells. For preliminary experiments, 1251-mEGF was of the supernatant. The WGL resin was eluted an additional covalently cross-linked to the EGF receptor in A431 cell seven times in rapid succession using the same procedure, membrane vesicles with SSFSB as described elsewhere (30). and eluates were pooled. For sequencing studies, intact A431 cells were used as a Tryptic Digestion ofPurified EGF Receptor Cross-Linked to source of EGF receptor as follows: confluent A431 cells were '"I-mEGF. The WGL eluate was dialyzed overnight at 4TC in detached from T75 flasks with gentle agitation Spectrapor no. 2 membrane dialysis tubing (Mr cutoff, for 30 min at 12,000-14,000) that had been pretreated by boiling for 20 mi room temperature with 5 ml of trypsin/EDTA (lx) per flask. in Milli-Q-filtered water (Millipore), against 4 liters of Milli- Harvested cells were pelleted by centrifugation at maximum Q-filtered water with one change of water. Aliquots of the speed for 5 min in a Sorvall GLC-1 centrifuge maintained at dialyzed material were added repeatedly to a 1-ml Reacti-Vial 40C and supernatants were discarded; the yield was -150 ,ul (Pierce) and dried in a Speed-Vac concentrator (Savant). of packed cells per flask. Cells were washed twice by EGF receptor cross-linked to 125I-mEGF, visible as an off- resuspension in 30 vol of an ice-cold solution of 20 mM white residue at the tip ofthe vial, was dissolved in 100 td of Hepes, pH 8.0/10mM iodoacetic acid/i mM EGTA/0.1 mM a solution of 8 M urea/0.4 M ammonium bicarbonate and phenylmethylsulfonyl fluoride, followed by centrifugation as reduced and carboxyamidomethylated as described (36). described above and discarding of supernatants. Washed Digestion with trypsin was carried out with stirring at 3TC in A431 cells were stored at -70°C or used immediately. a vol of 0.4 ml, with a total of 4 pug of trypsin added in three 17-5I-mEGF (specific activity, 30,000 cpm/,ug) was deriva- aliquots over 24 h. Digestion was halted with the addition of tized with SSFSB and separated from excess reagent by gel a 12x molar excess of soybean trypsin inhibitor. filtration as described (30) and applied at a final concentration Immunoprecipitation of EGF-Linked Tryptc Peptides. The of 0.12 ,uM to 4-5 ml of washed A431 cells suspended in 160 tryptic digest was transferred to a 5-ml glass tube, and 0.6 ml ml of ice-cold buffer containing 50 mM Hepes (pH 8.0), 0.1 of immunoprecipitation buffer (50 mM TrisHCI, pH 8.0/ mM Na3VO4, 5 mM MgCl2, 1 mM MnCl2, and 40 p.M ATP. 0.02% SDS/150 mM NaCl), 0.4 ml of anti-EGF immune Cross-linking of 1251-mEGF to the EGF receptor and receptor antiserum, and 0.2 ml of hydrated, packed PAS resin were autophosphorylation were allowed to proceed with gentle added. The resulting slurry was rocked overnight at 4TC. The agitation for 4 h at 4°C, with addition of 6.4 ,umol of PAS resin was pelleted for 10 min at 40C in the GLC-1 concentrated fresh ATP at 3.5 h. Centrifugation (27,000 x g) centrifuge as described above for other resins, and the for 20 min at 4°C followed, the supernatant was discarded, supernatant was removed; the beads were then washed six and pelleted cells were solubilized by alternately pipetting times by resuspension in 0.8 ml of immunoprecipitation and Vortex mixing them at 4°C in an ice-cold solution of 40 buffer, followed by centrifugation and removal of superna- ml of 20 mM Hepes (pH 7.4), 5% Triton X-100, 10% (vol/vol) tants as described above for other resins. To immunoprecip- glycerol, 1 mM EGTA, 0.1 mM Na3VO4, 5 mM MjCl2, 1 mM itate additional 12-5I-mEGF-linked peptide, the first four su- MnCl2, and 100 pM ATP. Solubilized cells were clarified by pernatants obtained in the procedure described above were centrifugation (213,000 x g) for 20 min at 4°C. pooled, and 0.4 ml of hydrated, packed PAS resin and 0.3 ml Anti-Phosphotyrosyl Affinity Purification of EGF Receptor of anti-EGF immune antiserum were added. The resulting Covalently Linked to 12'I-mEGF. The supernatant from la- slurry was rocked for 3.5 h at 40C before pelleting and beled, solubilized A431 cells was divided into four aliquots, washing six times as described above with 3 ml of immuno- and each aliquot was applied to a hydrated vol of 1 ml of APY precipitation buffer per wash. resin moistened by APY buffer [20mM Hepes (pH 7.4), 0.2% Electrophoresis of Anti-EGF Immunopreciptates. Electro- Triton X-100, 10%o glycerol, 1 mM EGTA, and 0.1 mM phoresis in SDS/Tricine gels was performed as described (37) Na3VO4]. Adsorption of receptor to the resin was accom- using gel phases with the following composition [the nomen- plished with gentle rocking of the slurry for 2 h at 4°C. clature is that of Hjerten (38)]: stacking phase, 4% T, 3% C; Centrifugation of the slurry in the same manner as described spacer phase, 10%1 T, 3% C; separating phase, 16.5% T, 6% above for A431 cells was followed by washing the pellet four C. Glycerol (13%) was present in the separating phase. Gels times by resuspension in APY wash buffer (APY buffer were preelectrophoresed with 7.5 mg of mercaptoacetic acid. supplemented with 150 mM NaCl, 5 mM MgCl2, 1 mM To the 0.6 ml of washed, combined PAS beads was added an MnCl2, and 0.1 mM ATP); centrifugation; and discarding of equal volume of SDS-containing 2x electrophoresis buffer supernatants as described above. Labeled receptor was (as in ref. 37, except that 100 mM dithiothreitol was used as eluted by rocking the resin overnight at 4°C in 2 ml of APY the reductant), and this suspension was subjected to electro- wash buffer supplemented with 150 mM NaCl, 5 mM phenyl phoresis at 100 V for 18 h. An overnight autoradiograph was phosphate, and 0.05% sodium azide. The resin was pelleted obtained ofthe wet gel mounted on Whatman 3 mm paper and as described above, and the supernatant was collected as sealed between plastic sheets. eluate. The resin was resuspended in another 2 ml of elution Elution of Immunoprecipitated Peptides from Polyacrylam- buffer, centrifugation and supernatant collection were re- ide Gels. Using the autoradiograph as a guide, the region of peated immediately, and eluates were pooled. the wet gel containing the labeled peptide of interest was Downloaded by guest on September 28, 2021 Biochemistry: Woltjer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 7803 excised, and the sample of wet gel was homogenized in 5 ml with control lanes containing undigested 125I-mEGF-linked of Milli-Q water. The minced acrylamide was incubated with receptor, undigested 1751I-mEGF, and digested 125I-mEGF to agitation for 4 h at 40C and then pelleted for 10 min at 40C as assay for digest completeness. The presence ofspecies in the described above for resins. The supernatant was retained, digest of "251-mEGF-linked receptor that could be attributed and peptide elution with 5 ml of water was repeated as to a tryptic fragment of I251-mEGF covalently linked to a described above, but with overnight incubation of the resus- tryptic fragment(s) of the EGF receptor was determined in pended acrylamide pellet. Pooled eluates were concentrated the following manner. The mI radiolabel in '25I-mEGF is to 0.1 ml of viscous liquid using a Speed-Vac evaporator. To carried on tyrosyl residues; since all cleavage sites for trypsin this liquid was added 0.15 ml of methanol, and the resulting are C-terminal to all tyrosyl residues, a complete tryptic mixture was added in aliquots of 75 A1 to a protein support digest of 125I-mEGF can be expected to give rise to a single disc (Porton Instruments, Tarzana, CA) in a 5-ml glass tube. radiolabeled peptide derived from the N-terminal portion of After addition of each aliquot, the disc was dried for 30 min 5I-mEGF (Fig. 1). Furthermore, since cross-linking pro- in a Speed-Vac concentrator, washed by agitation for 1 min ceeds through the N terminus of 125I-mEGF, linkage to a with 1 ml of methanol followed by removal of methanol by single EGF receptor peptide would likewise be expected to pipetting, and redried as described above. After application give rise to a unique radiolabeled peptide. of all aliquots, two additional methanol washes were per- The results of digests with trypsin (Fig. 2) approximate formed as described above, except that the disc was left to these expectations. Trypsin-cleaved 125I-mEGF was visual- soak in methanol for 1 h at room temperature before removal ized on autoradiographs as a diffuse band that migrated more of methanol. slowly than intact '25I-mEGF. This anomalous migration may Sequendng of EGF-Linked Peptides. Sequencing of Porton be due to the acidic nature of mEGF and the loss of disc-supported samples was performed on an Applied Bio- hydrophobic residues upon cleavage with trypsin. A radio- systems model 475A sequencer. Phenylthiohydantoin (PTH) labeled species comigrating with trypsin-digested 125I-mEGF amino acid derivatives were separated on a model 120A was present in the tryptic digest of EGF receptor that had on-line analyzer by the column and separation protocol been cross-linked to 125I-mEGF; this is expected, due to the provided by the manufacturer. Chromatographic data were presence of the bound but unlinked 125I-mEGF carried collected and analyzed with an Applied Biosystems 900A through the receptor purification described above. Addition- data controller with the supplied data acquisition software. ally present in the digest of 125I-mEGF-linked receptor, however, was a species with approximately the same migra- tion as intact l251-mEGF. It was concluded that this uniquely RESULTS AND DISCUSSION migrating band is likely to represent a receptor-derived In a previous paper, we reported the development of the peptide linked to the N terminus of trypsin-treated "-'I- technique used here to achieve specific, covalent attachment mEGF. of SSFSB-modified 125I-mEGF to >60%o of specific binding Attempts to transfer this peptide electrophoretically to a sites for EGF through the a-amino group of 125I-mEGF (30). poly(vinylidene difluoride) sequencing membrane (40) gave To contribute to understanding the sites of interaction of rise to poor yields of transferred protein, probably because EGF with its receptor, we have isolated and determined the the hydrophobic residues in mEGF that may mediate binding amino acid sequence of a tryptic peptide containing the site to poly(vinylidene difluoride) are lost after tryptic digestion. in the EGF receptor to which the N terminus of 1251-mEGF Moreover, multiple PTH amino acid derivatives were de- had been cross-linked. tected in each sequencing cycle when a poly(vinylidene As described above, 1251-mEGF was affinity cross-linked difluoride) membrane containing a small amount of bound to EGF receptor-rich A431 cells. Under the conditions de- peptide was subjected to sequence analysis (unpublished scribed, the cross-linking of radiolabeled ligand to the recep- data). It was apparent that a single electrophoretic step was tor is essentially complete within 4 h (39). Experiments with not sufficient to separate the products of the tryptic digest shed membrane vesicles from A431 cells showed that the and that other media for peptide adsorption for sequencing EGF cross-linked to the receptor in this manner stimulates were indicated. autophosphorylation of the receptor at least as well as an equivalent amount of free mEGF (30). Triton-solubilized EGF receptor cross-linked to 1251-mEGF could be purified to -80% homogeneity in a single affinity chromatographic step, as assayed by silver staining of SDS gels of APY eluates, with recovery of >70%o of receptor-linked radioactivity (39). For the purpose of removing excess Triton from the sample, APY eluates were applied to WGL resin, and, after washing the resin in detergent-free buffer, >70o of the applied radioactivity was successfully eluted. Electrophore- sis of WGL eluates in SDS gels revealed that approximately half of the eluted radioactivity comigrated with the EGF receptor, and the rest comigrated with free EGF. Control experiments (unpublished data) indicated that the covalent cross-link of 1251-mEGF to the EGF receptor is as stable as the covalent integrity ofthe EGF receptor itself; hence, it was concluded that the presence of free 1251-mEGF in eluates even after extensive purification of the EGF receptor is due 0 to binding of 125I-mEGF to the receptor without covalent attachment. After dialysis for the purpose of removing salts FIG. 1. Modifications present in mEGF, which was affinity and glycerol, 98% of the radioactivity present in the WGL cross-linked to the EGF receptor. The N terminus of SSFSB- eluate was recovered in the dialysate and concentrated. modified mEGF bears the fluorosulfonylbenzoyl moiety through EGF receptor cross-linked to 125I-mEGF was reduced, which cross-linking to the EGF receptor occurs. Tyrosyl residues (*) carboxyamidomethylated, and subjected to tryptic digestion. carry the radiolabel in MI-mEGF. Sites of tryptic cleavage of An aliquot of the digest was fractionated by electrophoresis, reduced carboxyamidomethylated mI-mEGF are denoted by t. Downloaded by guest on September 28, 2021 7804 Biochemistry: Woltjer et al. Proc. Nad. Acad. Sci. USA 89 (1992)

A 2 3 4 3 1 2 of the receptor, by comparison to a tryptic cleavage map of the cDNA-derived receptor sequence. Under conditions of 3 complete digest, this peptide would be predicted to contain 21 -F:. -'I receptor residues, in addition to 41 residues from linked, trypsin-treated I251-mEGF. PTH-derivatized amino acids at- tributable to the amino acid sequence of mEGF were not detected, lending further support to earlier work (30), which --IF-(,1t suggested that cross-linking occurred through the terminal amino group of 125I-mEGF. Lysyl, tyrosyl, cysteinyl, and histidyl residues are the most likely targets in the receptor for reaction with 175I-mEGF FIG. 2. Tricine gel electrophoresis of the products of tryptic bearing a reactive fluorosulfonylbenzoyl moiety. Linkages to digestion of mlI-mEGF-linked receptor. Samples ofintact or trypsin- lysyl or tyrosyl residues would be expected to be stable to digested 125I-mEGF or 125I-mEGF-linked receptor were prepared sample treatment with trifluoroacetic acid during the course and separated by electrophoresis as described in the text; the figure ofsequencing (41) and would give rise to chromatograms with consists of autoradiographs of unfixed, dried polyacrylamide gels. low levels ofthe corresponding PTH-derivatized amino acids A431 membrane vesicles were used as a source of receptor in A, in cycles in which these residues would be predicted which depicts intact 1251-mEGF-linked receptor (EGFR-EGF) with from the copurified u5I-mEGF (EGF) (lane 1), digested 125I-mEGF-linked receptor sequence. receptor (lane 2), intact 125I-mEGF (lane 3), and digested 125I-mEGF Fig. 4 shows that signals from tyrosyl residues are present (lane 4). The application of trypsin to lUI-mEGF-linked receptor at approximately the expected yields in sequencing cycles 4, under the conditions described is seen to generate two bands in the 5, and 9; these residues, then, appear not to have been sites low molecular weight region of the gel. The more slowly m ting of linkage to 1zI-mEGF. The signal for Tyr-101 of cycle 17, band comigrates with trypsin-digested 'zI-mEGF (tEGF), and the however, is much diminished in both experiments A and B. remaining band, which comigrates with intact 125I-mEGF, was attributed to a tryptic fragment of the EGF receptor linked to This lends support to the hypothesis that Tyr-101 is a site of trypsin-digested lzI-mEGF (tEGFR-EGF). InB, A431 cells were the linkage to 125I-mEGF; and the nearly quantitative absence of source of the purified, trypsin-treated 125I-mEGF-linked receptor of the signal for Tyr-101 in sequencing chromatograms implies lane 1, and two less-well-resolved bands can again be distinguished. that Tyr-101 is virtually the only site of linkage. The accel- The more quickly migrating band comigrates with the intact 'ZI- erated loss of sequenceable peptide in cycles after cycle 17 is mEGF present in lane 2. The sample from which the aliquot of lane consistent with this hypothesis, since sequencing cycle 17 1 was derived was further purified and sequenced successfully. would have resulted in cleavage of the remainder of the Serum containing antibody directed against native mEGF receptor firgment from w'I-mEGF, which may have helped was found also to immunoprecipitate reduced, carboxyami- tether the fragment to the sequencing support. domethylated, trypsin-treated 125I-mEGF (39). Hence, anti- 1.0 -... mEGF antibody and PAS were applied to the tryptic digest A of the purified preparation of EGF receptor linked to 1251- 0.5 mEGF; 49%o ofthe radioactivity present in the digest resided 0.0 in washed immunoprecipitates. Electrophoresis and autora- -- diography ofthe immunoprecipitate revealed the presence of o -0.5 both trypsin-treated 125I-mEGF and the species believed to o -1.0 be 125I-mEGF-linked EGF receptorfragment. Approximately '0 0 26 pmol of the latter species (as estimated from the radioac- 2 -1.5 la tivity present) was eluted from the gel, and 13 pmol ofeluate U2 was successfully applied to a Porton protein support disc. 2 -2.0 '4.4 2 4 6 8 10 12 14 16 18 2 Do 1251-mEGF-linked peptides derived from two large-scale 0 affinity cross-linking experiments were sequenced; experi- *1 0.5 ment B was carried out by using essentially the same pro- 1-1 B cedures as in experiment A described above. Sequencing 10.0 laon chromatograms from both experiments indicated the pres- .0 0.5 0 ence ofa single peptide (Fig. 3), which was recognized as the .00 ~0.0 fragment produced by tryptic cleavage at Arg-84 in domain I *0

-0.5 1 Cycle # 1 17

Experiment A XXMYYXNSYALAVLXN-D -1.0

Experiment B XXMYYEXSYALAVLSX-DA -1.5 2 4 6 8 10 12 14 16 18 20 ReoeptorSeq. IRGNMYYENSYALAVLSNYDANKT Cycle Receptor Pos. 84 101 105 FIG. 4. Yields of PTH-derivatized amino acid derivatives ob- FIG. 3. Sequence analysis of a tryptic peptide of the EGF served during sequence analysis of a tryptic peptide of the receptor receptor affinity cross-linked to 125I-mEGF. Identified residues in affinity cross-linked to mI-mEGF. The background-subtracted two experiments (A and B) are aligned with the corresponding yields of the residues identified in Fig. 3 are plotted as a function of portion of the cDNA-derived EGF receptor sequence. The sequenc- sequencing cycle number. Line represents least-squares fit of the ing cycle is given in the top line, and the positions ofrelevant receptor logarithm of the yields to the cycle number, where cycle 17 was residues are in the bottom line. X, sequencing cycles for which excluded from the fit calculation. (A) Results from experiment A. (B) residues could not be identified due to contaminants in chromato- Results from experiment B described in the text. Initial sequencing grams, which precluded identification and quantitative evaluation of yields were estimated to be 509O, and apparent repetitive yields residues. Dashes indicate sequencing cycles in which a residue could were -85%. PTH-derivatized amino acids were indistinguishable not be identified despite the absence ofcontaminant peaks at or near from background noise after cycle 18 in experiment A and after cycle the anticipated residue in sequencing chromatograms. 19 in experiment B. Downloaded by guest on September 28, 2021 Biochemistry: Woltjer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 7805 The cDNA-derived receptor sequence shows that Tyr-101 1. Cohen, S. (1960) Proc. Natl. Acad. Sci. USA 46, 302-311. is followed by the tryptic cleavage sites Lys-105, Lys-109, 2. Cohen, S. (1962) J. Biol. Chem. 237, 1555-1562. 3. Carpenter, G. & Wahl, M. I. (1990) Handb. Exp. Pharmacol. 95,69-171. and Arg-114; that no cysteinyl or histidyl residues exist 4. Boni-Schnetzler, M. & Pilch, P. F. (1987) Proc. Natd. Acad. Sci. USA 84, within residues 85-114; and that no additional tyrosyl resi- 7832-7836. dues are located within residues 102-114. Therefore, if Tyr- 5. Yarden, Y. & Schlessinger, J. (1987) Biochemistry 26, 1434-1442. 101 is not the site of linkage to 125I-mEGF, Lys-105 and 6. Yarden, Y. & Schlessinger, J. (1987) Biochemistry 26, 1443-1451. Lys-109 would be the next most likely candidates. 7. Cochet, C., Kashles, O., Chambaz, E. M., Borrello, I., King, C. R. & Schlessinger, J. (1988) J. Biol. Chem. 263, 3290-3295. In two previous cross-linking studies (26, 27), methods that 8. Northwood, I. C. & Davis, R. J. (1988) J. Biol. Chem. 263, 7450-7453. did not involve direct sequencing were used to identify 9. Fanger, B. O., Stephens, J. E. & Staros, J. V. (1989) FASEBJ. 3, 71-75. domain III of the EGF receptor as containing sites oflinkage 10. Carpenter, G., King, L., Jr., & Cohen, S. (1979) J. Biol. Chem. 254, to 125I-mEGF. With the result reported here, the results from 4884-4891. 11. Ushiro, H. & Cohen, S. (1980) J. Biol. Chem. 255, 8363-8365. cross-linking experiments as a whole complement and add 12. Buhrow, S. A., Cohen, S. & Staros, J. V. (1982) J. Biol. Chem. 257, detail to studies of receptor mutants, which imply roles for 4019-4022. both domains I and III in the binding of EGF to its receptor. 13. Buhrow, S. A., Cohen, S., Garbers, D. L. & Staros, J. V. (1983) J. Biol. It should be noted that affinity cross-linking is essentially Chem. 258, 7824-7827. always an "exo" labeling technique in the terminology of 14. Ullrich, A., Coussens, L., Hayflick, J. S., Dull, T. J., Gray, A., Tam, A. W., Lee, J., Yarden, Y., Libermann, T. A., Schlessinger, J., Down- Baker (42)-i.e., with the labeling reaction occurring outside ward, J., Mayes, E. L. V., Whittle, N., Waterfield, M. D. & Seeburg, of the binding site itself. P. H. (1984) Nature (London) 369, 418-424. That the cross-linking reagent used in other studies, dis- 15. Cohen, S., Ushiro, H., Stoscheck, C. & Chinkers, M. (1982) J. Biol. uccinimidyl suberate (DSS) (43), which was also expected to Chem. 257, 1523-1531. mediate cross-linking through the N 16. Gill, G. N., Bertics, P. J. & Santon, J. B. (1987) Mol. Cell. Endocrinol. terminus of 125I-mEGF, 51, 169-186. was longer and more flexible than the SSFSB used here may 17. Staros, J. V., Fanger, B. O., Faulkner, L. A., Palaszewski, P. P. & account for the different results obtained for sites of cross- Russo, M. W. (1989) in Receptor Phosphorylation, ed. Moudgil, V. K. linking to the receptor. Moreover, in the direct cross-linking (CRC, Boca Raton, FL), pp. 227-242. experiments reported elsewhere, the possibility of initial 18. Ullrich, A. & Schlessinger, J. (1990) Cell 61, 203-212. 19. Murthy, U., Basu, A., Rodeck, U., Herlyn, M., Ross, A. H. & Das, M. modification by DSS of receptor sites that are readily acces- (1987) Arch. Biochem. Biophys. 252, 549-560. sible to the cross-linking reagent, followed by cross-linking to 20. Wu, D., Wang, L., Sato, G. H., West, K. A., Harris, W. R., Crabb, residues of 125I-mEGF other than its N terminus, cannot be J. W. & Sato, J. D. (1989) J. Biol. Chem. 264, 17469-17475. excluded. DSS could be expected to display some reactivity 21. Defize, L. H., Boonstra, J., Meisenhelder, J., Kruijer, W., Tertoolen, toward, for example, tyrosyl residues of 125I-mEGF that L. G. J., Tilly, B. C., Hunter, T., van Bergen en Henegouwen, P. M. P., Moolenaar, W. H. & de Laat, S. W. (1989) J. Cell Biol. 109, 2495-2507. could be present in high local concentration at the site of 22. Bellot, F., Moolenaar, W., Kris, R., Mirakhur, B., Verlaan, I., Ullrich, receptor modification by DSS. The thorough characteriza- A., Schlessinger, J. & Felder, S. (1990) J. Cell Biol. 110, 491-502. tion of cross-linker-modified mEGF in our studies (30), 23. Lax, I., Fischer, R., Ng, C., Segre, J., Ullrich, A., Givol, D. & however, eliminates the possibility of cross-linking at sites Schlessinger, J. (1991) Cell Regul. 2, 337-345. other than the N terminus of 125I-mEGF. 24. Lax, I., Bellot, F., Howk, R., Ullrich, A., Givol, D. & Schlessinger, J. (1989) EMBO J. 8, 421-427. The affinity cross-linking described here was performed 25. Lax, I., Bellot, F., Honegger, A., Schmidt, A., Ullrich, A., Givol, D. & with membrane-resident, intact EGF receptor. The most Schlessinger, J. (1990) Cell Regul. 1, 173-188. specific cross-linking result published to date (27), as well as 26. Lax, I., Burgess, W. H., Bellot, F., Ullrich, A., Schlessinger, J. & Givol, all electron microscopic and x-ray crystallographic imaging D. (1988) Mol. Cell. Biol. 8, 1831-1834. 27. Wu, D., Wang, L., Chi, Y., Sato, G. H. & Sato, J. D. (1990) Proc. Nail. of the EGF receptor's ligand binding domain, were derived Acad. Sci. USA 87, 3151-3155. from work with secreted or truncated receptor forms. Al- 28. Lax, I., Mitra, A. K., Ravera, C., Hurwitz, D. R., Rubinstein, M., though studies with conformation-sensitive antibodies have Ullrich, A., Stroud, R. M. & Schlessinger, J. (1991) J. Biol. Chem. 266, failed to detect differences between these species and the 13828-13833. extracytoplasmic portion of the intact receptor (27), it is 29. Gunther, N., Betzel, C. & Weber, W. (1990) J. Biol. Chem. 265, 22082-22085. known that the intact, unsolubilized receptor binds 125I- 30. Woltjer, R. L., Weclas-Henderson, L., Papayannopoulos, I. A. & Sta- mEGF with 100-fold higher affinity than the secreted or ros, J. V. (1992) Biochemistry 31, in press. truncated ligand-binding portion (28, 44, 45). It is conceivable 31. Savage, C. R., Jr., & Cohen, S. (1972) J. Biol. Chem. 247, 7609-7611. that differences in the affinity with which 125I-mEGF is bound 32. Carpenter, G. & Cohen, S. (1976) J. Cell Biol. 71, 159-171. to its receptor could be reflected in differences in receptor 33. Frackelton, A. R., Jr. (1983) Cells 3, 339-345. 34. Wahl, M. I., Daniel, T. 0. & Carpenter, G. (1988) Science 241, 968-970. residues in proximity to the N terminus of cross-linker- 35. Frackelton, A. R., Jr., Ross, A. & Eisen, H. (1983) Mol. Cell. Biol. 3, derivatized 12-I-mEGF. Interestingly, a recent report indi- 1353-1360. cates that DSS-mediated cross-linking of 125I-mEGF to the 36. Stone, K. L., LoPresti, J. B., Crawford, J. B., DeAngelis, R. & Wil- EGF receptor may not occur in all preparations of the liams, K. R. (1989) in A Practical Guide to Protein and Peptide Purifi- extracytoplasmic of cation for Microsequencing, ed. Matsudaira, P. T. (Academic, San portion the receptor (46). Diego), pp. 31-74. The cross-linking methods described here enabled the first 37. Schlgger, H. & von Jagow, G. (1987) Anal. Biochem. 166, 368-379. elucidation, by direct protein microsequencing, of a site of 38. Hjerten, S. (1962) Arch. Biochem. Biophys. 99, 466-467. interaction of EGF and the EGF receptor. It is hoped that 39. Woltjer, R. L. (1992) Dissertation (Vanderbilt Univ., Nashville, TN). these techniques have further potential for contributing to our 40. Matsudaira, P. (1987) J. Biol. Chem. 262, 10035-10038. 41. Annamalai, A. E. & Colman, R. F. (1981) J. Biol. Chem. 256, 10276- understanding ofhow EGF interacts with its receptor, and for 10283. laying a foundation for the asking ofdetailed questions about 42. Baker, B. R. (1967) Design ofActive-Site-Directed Irreversible Enzyme the functions of EGF binding. Inhibitors (Wiley, New York). 43. Pilch, P. F. & Czech, N. P. (1979) J. Biol. Chem. 254, 3375-3381. We acknowledge the assistance of U. Barnela in the preparation 44. Basu, A., Raghunath, M., Bishayee, S. & Das, M. (1989) Mol. Cell. Biol. 9, 671-677. and radiolabeling of mEGF. We are grateful to D. Sanchez for her 45. Greenfield, C., Hiles, I., Waterfield, M. D., Federwisch, M., Wollmer, work with A431 cell cultures. We thank Dr. L. Weclas-Henderson A., Blundell, T. L. & McDonald, N. (1989) EMBO J. 8, 4115-4123. for the preparation of SSFSB. This work was supported by Grants 46. Rubinstein, M., Felder, S., Lax, I., Zhou, M., Ullrich, A. & Schlessinger, R01 DK25489, R01 GM30861, T32 GM08320, T32 GM07347, and J. (1992) FASEB J. 6, A94 (abstr.). 2 S07 RR05424-30-41 from the National Institutes of Health. 47. Woljer, R. L. & Staros, J. V. (1990) FASEB J. 4, A2208 (abstr.). Downloaded by guest on September 28, 2021