Direct linkage of EGF to its receptor: characterization and biological relevance

Preclinical modeling of EGFR inhibitor resistance in head and neck cancer

The epidermal growth factor receptor (EGFR) is widely expressed in head and neck squamous cell carcinomas (HNSCC) and can activate many growth and survival pathways within tumor cells. Despite ubiquitous EGFR expression, therapies targeting the receptor are only modestly effective in the treatment of HNSCC. A consistent mechanism of resistance to EGFR targeting agents has not yet been identified in HNSCC likely due, in part, to the paucity of preclinical models. We assessed the in vitro and in vivo responses of a panel of 10 genotypically validated HNSCC cell lines to the EGFR inhibitors erlotinib and cetuximab to determine their validity as models of resistance to these agents. We defined a narrow range of response to erlotinib in HNSCC cells in vitro and found a positive correlation between EGFR protein expression and erlotinib response. We observed cross-sensitivity in one HNSCC cell line, 686LN, between erlotinib and cetuximab in vivo. We attempted to generate models of cetuximab resistance in HNSCC cell line-derived xenografts and heterotopic tumorgrafts generated directly from primary patient tumors. While all 10 HNSCC cell line xenografts tested were sensitive to cetuximab in vivo, heterotopic patient tumorgrafts varied in response to cetuximab indicating that these models may be more representative of clinical responses. These studies demonstrate the limitations of using HNSCC cell lines to reflect the heterogeneous clinical responses to erlotinib and cetuximab, and suggest that different approaches including heterotopic tumorgrafts may prove more valuable to elucidate mechanisms of clinical resistance to EGFR inhibitors in HNSCC.

Aberration of epidermal growth factor receptor expression in bone and soft-tissue tumors: protein overexpression, gene amplification and activation of downstream molecules

In order to evaluate the involvement of epidermal growth factor receptor, and to analyze the correlation between gene aberration and protein expression in mesenchymal tumors, we examined protein expression by immunohistochemistry in 125 cases of bone and soft-tissue tumors. Furthermore, amplification of epidermal growth factor receptor gene was determined by fluorescence in situ hybridization. Positive immunostaining was found in 23 cases (18.4%). Among these 23 cases, one of malignant fibrous histiocytoma showed the highest degree (3+) of protein overexpression and gene amplification as clusters of hybridization signals, indicating homogeneously staining regions. The second case of malignant fibrous histiocytoma also showed a higher degree (2+) of overexpression and coamplification of the epidermal growth factor receptor gene with the centromeric regions, indicating polysomy of chromosome 7. The levels of expression observed in immunohistochemistry were confirmed by immunoblotting and found to be comparable. Moreover, although expression of phosphorylated epidermal growth factor receptor was detected in those two cases of malignant fibrous histiocytoma, constitutive activation of extracellular signal-related protein kinase 1/2 was not observed, suggesting that activation of epidermal growth factor receptor does not necessarily and constantly lead to signal transduction to the downstream molecules. In the remaining 123 cases, including 21 cases exhibiting weak (1+) immunoreactivity, no gene amplification nor polysomy was found. Collectively, expression of epidermal growth factor receptor was observed not infrequently in mesenchymal tumors, but 'overexpression' is rare and can be attributed to an increase in gene copy number, resulting from amplification or polysomy. Although cases that scored positive for protein expression and/or gene amplification could be qualified candidates for antiepidermal growth factor receptor therapies, further examination of the status of downstream molecules in the signal cascade, such as phosphorylated epidermal growth factor receptor and extracellular signal-related protein kinase 1/2, may be required as the process of therapeutic strategy.

Src family kinases and HER2 interactions in human breast cancer cell growth and survival

Evidence from murine fibroblast models and human breast cancer cells indicates that c-Src and human EGF receptor (HER1) synergize to enhance neoplastic growth of mammary epithelial cells. To investigate whether interactions between c-Src and other HER family members may also play a role in breast tumor progression, we characterized 13 human breast carcinoma cell lines and 13 tumor samples for expression of HER family members and c-Src and examined a subset of the cell lines for Src-dependent, heregulin (HRG)-augmented, anchorage-dependent and independent growth. By immunoblotting, we found that all cell lines overexpressed one or more HER family member, and 60% overexpressed c-Src. Seventy-five per cent of the tumor tissues overexpressed HER2, while 64% overexpressed c-Src. Colony formation in soft agar was enhanced by HRG in three of five cell lines tested, a response that correlated with the presence of a c-Src/HER2 heterocomplex. This result suggests that HRG may act through both HER2 and c-Src to facilitate anchorage-independent growth. In contrast, HRG had little effect on anchorage-dependent growth in any of the cell lines tested. PP1, a Src family kinase inhibitor, reduced or ablated HRG-dependent and independent soft agar growth or anchorage dependent growth, and triggered apoptosis in all cell lines tested. The apoptotic effect of PP1 could be partially or completely reversed by HRG, depending on the cell line. These results suggest that while Src family kinases may cooperate with HRG to promote the survival and growth of human breast tumor cells, they also function independently of HER2/HRG in these processes.

Perturbation of epidermal growth factor clearance after radioiodination and its implications

The clearance of human epidermal growth factor (hEGF1-53) has been thought to be mediated mainly by a high-capacity receptor system, yet relatively low in vivo clearance rates (<10 mL/min/kg) and long terminal elimination half-lives (>120 min) have been observed in rats receiving the peptide that was iodinated by the oxidative chloramine-T (CT) method. We investigated if a mild, less oxidative iodination by the lactoperoxidase (Enzymobeads, EB) method, which is known to yield an iodinated peptide with receptor-binding equivalence, could produce a labeled peptide that behaves pharmacokinetically similar to the native material. For comparison, a parallel study was also conducted with EB-125I-hEGF1-48, which in its native form has a much reduced receptor binding activity due to the loss of the C-terminal pentapeptide. Plasma radioactivity concentrations were determined by trichloroacetic acid (TCA) precipitation and immunoprecipitation. Rats cleared unlabeled hEGF1-53 and hEGF1-48 markedly faster (CL(tot) > 120 mL/min/kg) than their radiolabeled counterparts. Approximately 96% of the hEGF1-53 dose was cleared during the initial phase (0-4 min), as opposed to only 5-14% for the iodinated peptide. Similar change was also observed for EB-125I-hEGF1-48 and CT-125I-hEGF1-53. The pharmacokinetic behavior of EB-125I-hEGF1-53 was, in fact, comparable to that of CT-125I-hEGF1-53. These observations indicate that receptor-binding equivalence does not have direct relationship with in vivo EGF clearance. Both iodination methods (oxidative CT and less oxidative EB) might have perturbed one or more steps in the cascade of ligand-receptor internalization and intracellular procession, which in turn modified the disposition of the peptides. In addition, the two independent precipitation techniques for the same peptide generated different kinetic outcomes. The overall experimental results suggest that it is unacceptable to use an iodinated form to characterize the disposition of peptides/proteins like EGF with a specific receptor system mediating its clearance.

Domain deletion in the extracellular portion of the EGF-receptor reduces ligand binding and impairs cell surface expression

Cultured NIH-3T3 cells were transfected with cDNA constructs encoding human epidermal growth factor-receptor (EGF-R)* and two deletion mutants in the extracellular portion of the receptor molecule. One mutant is devoid of 124 amino-terminal amino acids, and the other lacks 76 residues. Mutant receptors were not delivered to the cell surface unless the transfected cells contained also endogenous EGF-Rs, suggesting that receptor interaction complements the mutation and allows surface display of mutant receptors. Immunoprecipitation experiments revealed an association between mutant and endogenous EGF-Rs when both proteins were expressed in the same cell. Hence, receptor-oligomers may exist in the plane of the membrane even in the absence of ligand binding, and oligomerization may play a role in normal trafficking of EGF-Rs to the cell surface. Mutant receptors retained partial ligand binding activity as 125I-labeled EGF was covalently cross-linked to both mutant receptors, and EGF stimulated, albeit weakly, their protein tyrosine kinase activity. Both mutant EGF-Rs bind EGF with a 10-fold lower affinity than that of the solubilized wild type EGF-R. These results provide further evidence that the region flanked by the two cysteine-rich domains plays a crucial role in defining ligand-binding specificity of EGF-R.

Effect of oxidative iodination of epidermal growth factor on its binding and secretion by hepatocytes

Experiments were undertaken to determine whether the method of iodination of epidermal growth factor (EGF) affects its binding to rat liver plasma membranes and its uptake, processing, and secretion into bile by intact rat hepatocytes. EGF was iodinated using one of three oxidative reagents: chloramine T (CT), lactoperoxidase (LP), or monochloride (MC). Quantitative receptor binding studies on plasma membranes isolated from male rat livers with either CT-, LP-or MC-125I-EGF indicated no significant difference in the apparent binding constants of the three preparations. To determine whether these three preparations were capable of forming a covalent-like complex with the EGF receptor, they were individually incubated with isolated plasma membranes and subjected to polyacrylamide gel electrophoresis under reducing conditions, followed by autoradiography. Each preparation formed a major radioactive protein band of approximately 180 kD, identified as the EGF receptor by immunoprecipitation with monoclonal anti-EGF receptor antibodies. Furthermore, even unlabeled EGF incubated with plasma membranes formed this same 180 kD band, as revealed on Western blots using anti-EGF antibody. The biliary secretion of CT-, LP-, and MC-125I-EGF was compared by injecting each one into rat portal veins and measuring the total and immunoprecipitable radioactivity in bile. The amount of immunologically intact CT-125I-EGF in bile was significantly greater than the others, whereas MC-125I-EGF transport was significantly reduced. We conclude that the method of iodination does not affect the covalent-like binding properties of EGF. Furthermore, since unlabeled EGF displayed these same binding properties, oxidative iodination procedures per se do not account for the covalent-like association between EGF and its receptor. However, the method of iodination used did affect the intracellular transport and processing of EGF by hepatocytes. The structural modification responsible for this alteration in transport properties has yet to be determined.

Localization of a major receptor-binding domain for epidermal growth factor by affinity labeling

Epidermal growth factor (EGF) receptor was affinity labeled with 125I-labeled EGF, using bifunctional covalent cross-linking agents. The affinity-labeled receptor was isolated and cleaved with CNBr to yield a single-labeled fragment, which was unequivocally identified by site-specific antibodies and other methods to encompass residues 294 to 543 of the EGF receptor. On the basis of amino acid sequence conservation, the extracellular portion of EGF receptor can be divided into four domains. The labeled CNBr fragment contains the entire sequence which is flanked by the two cysteine-rich domains of extracellular portion of the EGF receptor denoted as domain III. On the basis of these and other results, we propose that domain III contributes most of the interactions that define ligand-binding specificity of the EGF receptor.

Localization of a major receptor-binding domain for epidermal growth factor by affinity labeling

Epidermal growth factor (EGF) receptor was affinity labeled with 125I-labeled EGF, using bifunctional covalent cross-linking agents. The affinity-labeled receptor was isolated and cleaved with CNBr to yield a single-labeled fragment, which was unequivocally identified by site-specific antibodies and other methods to encompass residues 294 to 543 of the EGF receptor. On the basis of amino acid sequence conservation, the extracellular portion of EGF receptor can be divided into four domains. The labeled CNBr fragment contains the entire sequence which is flanked by the two cysteine-rich domains of extracellular portion of the EGF receptor denoted as domain III. On the basis of these and other results, we propose that domain III contributes most of the interactions that define ligand-binding specificity of the EGF receptor.

Translocation of epidermal growth factor to the hepatocyte nucleus during rat liver regeneration

To determine the fate of exogenous epidermal growth factor (EGF) in regenerating liver, 125I-labeled EGF was injected into rat portal veins at various times after 70% hepatectomy. Epidermal growth factor was taken up by the liver remnant at all time points studied (0, 4, 8, 16, and 36 h), but at 8 h after hepatectomy a large quantity was retained by the liver and EGF degradation products appearing in the bile decreased markedly. Electron microscopic autoradiography of the regenerating livers 1 h after injection of 125I-EGF demonstrated that 27% of the grains were associated with hepatocyte nuclei compared to 0.5% in shamoperated controls. There was also a concomitant decrease in grains associated with the lysosomal compartment. Nuclei isolated from regenerating livers exposed to 125I-EGF also demonstrated a three-fold increase in radioactivity compared to nuclei from control livers. Nearly 70% of nuclear radioactivity was precipitable with a specific antibody to EGF, and a small fraction appeared to be part of a high molecular weight complex. These data support the hypothesis that during the pre-S phase of liver regeneration, EGF is translocated to the nucleus rather than to lysosomes, and may participate in the initiation of deoxyribonucleic acid synthesis or alteration of gene expression.

Epidermal growth factor induces rapid, reversible aggregation of the purified epidermal growth factor receptor

Epidermal growth factor (EGF) receptor from A-431 cells was purified by affinity chromatography with monoclonal anti-receptor antibodies. The purified radiolabeled receptor was incubated with EGF and then analyzed by gel electrophoresis under nondenaturing conditions. In these gels, the EGF receptor migrates in two forms: a fast-migrating (low) form and an EGF-induced slow-migrating (high) form. On the basis of the various control and calibration experiments described, it is concluded that the low form represents the monomeric 170-kilodalton EGF receptor and the high form represents an EGF receptor dimer. The binding of EGF causes a rapid, temperature-sensitive dimerization of the EGF receptor. Receptor dimerization is fully reversible and involves saturable, noncovalent interactions that are stable at neutral pH and in nonionic detergents. Both the monomeric and dimeric forms of the receptor bind EGF and undergo self-phosphorylation. The dimeric form of the receptor may possess higher ligand binding affinity, and it seems to be phosphorylated earlier than the monomeric form following the addition of EGF and [gamma-32P]ATP. On the basis of these results, it is concluded that receptor oligomerization is an intrinsic property of the occupied EGF receptor and that it may play a role in the activation of the kinase function and the subsequent transmembrane signaling process.

[Epidermal growth factor: structure, location, phosphorylation and regulation of its receptor]

Epidermal growth factor (EGF) is a Mr 6045 polypeptide first characterized for its ability to stimulate mitogenesis in epidermal and epithelial cells. The first step in the action of the growth factor is its binding to specific, high affinity membrane receptors. These receptors have been studied in a number of tissues and cell culture lines. The level of EGF receptors is modulated by many agents. EGF down-regulates its receptor. In addition, the number of EGF receptors is decreased by other growth factors (platelet-derived growth factor; transforming growth factor), by many tumor promoters and by viral transformation. Several hormones also can regulate EGF binding in its target tissues.

Progesterone receptor subunits are high-affinity substrates for phosphorylation by epidermal growth factor receptor

Purified preparations of epidermal growth factor (EGF) receptor were used to test hen oviduct progesterone receptor subunits as substrates for phosphorylation catalyzed by EGF receptor. Both the 80-kilodalton (kDa) (A) and the 105-kDa (B) progesterone receptor subunits were phosphorylated in a reaction that required EGF and EGF receptor. No phosphorylation of progesterone receptor subunits was observed in the absence of EGF receptor, even when Ca2+ was substituted for Mg2+ and Mn2+. Phospho amino acid analysis revealed phosphorylation at tyrosine residues, with no phosphorylation detectable at serine or threonine residues. Two-dimensional maps of phosphopeptides generated from phosphorylated 80- or 105-kDa subunits by tryptic digestion revealed similar patterns, with resolution of two major, several minor, and a number of very minor phosphopeptides. The Km of progesterone receptor for phosphorylation by EGF-activated EGF receptor was 100 nM and the Vmax was 2.5 nmol/min per mg of EGF receptor protein at 0 degrees C. The stoichiometry of phosphorylation/hormone binding for progesterone receptor subunits was 0.31 at ice-bath temperature and approximately 1.0 at 22 degrees C.

Covalent attachment of 125I-beta nerve growth factor to its receptors on sympathetic neurons

When 125I-beta nerve growth factor binds to sympathetic and sensory neurons, some labeled ligand is sequestered (becomes inaccessible to the external milieu) in a time- and energy-dependent manner. It would appear that the higher affinity receptor (type I) participates in this process to a greater extent than does the lower affinity receptor (type II) [ Olender and Stach , 1980; Olender et al., 1981]. A small portion of the sequestered 125I-beta nerve growth factor is found as part of a high molecular weight complex. When cells, which have been incubated with 125I-beta nerve growth factor, are solubilized with Triton X-100 and subjected to sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis, a complex with an apparent molecular weight of approximately 240,000 is obtained. The formation of the covalent complex can be prevented by the prior addition of excess unlabeled beta nerve growth factor or sodium fluoride and dinitrophenol. The covalent 125I-beta nerve growth factor-receptor complex is dissociated in 50 mM dithiothreitol indicating that disulfide linkages are involved. At concentrations of beta nerve growth factor (3.8 X 10(-11) -3.8 X 10(-10) M) where maximal fiber outgrowth occurs in vitro, approximately 50-266 attomoles (0.3-1.6% of the type I receptors) of the covalent complex are formed per 10(7) nerve cells. These data suggest that a small portion of the 125I-beta nerve growth sequestered by sympathetic neurons becomes covalently attached to its receptor subsequent to its sequestration in a manner which appears to involve type I receptors.

Molecular species of epidermal growth factor carrying immunosuppressive activity

The suppression of antibody formation to sheep red cells in mice by partially purified fractions of mouse submaxillary gland was shown to be caused by epidermal growth factor (EGF). Purification of EGF by the method of Savage and Cohen resolved three components referred to as EGF a, EGF b, and EGF c. All three induced premature eye opening in neonatal mice, but only EGF a (identified as EGF 1-53) had full immunosuppressive activity. EGF c was shown by micropeptide mapping of chymotryptic and thermolytic digests and amino-terminal analysis to differ from EGF a only by the presence of beta-aspartyl instead of an asparaginyl residue. EGF b differed from EGF a in that it lacked the N-terminal asparagine. EGF shortened enzymatically at its carboxy terminal by two or five amino acids did not have any immunosuppressive activity. These findings suggest that, in contrast to some other biological effects of EGF, intact amino and carboxy terminals are required for the expression of immunosuppressive activity.

Role of polypeptide growth factors in normal and abnormal growth

In the preceding sections we have shown evidence that growth-promoting factors are involved in three basic situations. In normal embryonic development and function of mature organisms, growth factors such as NGF and EGF are of prime importance in supporting the necessary embryonic cell proliferation and the development of specific cell types. Other factors operate on subsets of mature cells during specialized functions such as inflammation. Included in this set would be factors such as CSF/MGF and Interleukin-2. Another basic function of growth factors has been shown to be wound repair and organ regeneration. This includes the well characterized PDGF and FGF as well as the various renotropic factors and liver growth factors. As these factors must operate in mature organisms with many different cell types and similar cell types in many locations, more specificity is needed than in embryonic growth. This has resulted in the organ specific factors such as the renotropins and in the unique delivery system of the PDGF. The recent discovery and characterization of the transforming growth factors has provided a possible connection between embryonic and normal developmental growth and the rapid cellular proliferation characteristic of tumor cells. The TGF not only interacts with receptors for normal growth factors such as EGF but are also detectable in low levels in normal tissue and embryos. The exact relationships between these various factors will have to await the determinations of more amino acid sequences for comparisons. The other tumor-related product, tumor angiogenesis factor, is also found in normal tissue and inflammatory reaction sites.(ABSTRACT TRUNCATED AT 250 WORDS)

Regulation of cell proliferation by epidermal growth factor

Epidermal Growth Factor (EGF) is a 6045 dalton polypeptide which stimulates the proliferation of various cell types in vitro and in vivo. EGF binds to diffusely distributed membrane receptors which rapidly cluster primarily on coated pits areas on the plasma membrane. Subsequently, the EGF-receptor complexes are endocytosed and degraded by lysosomal enzymes. The lateral diffusion coefficient (D) of EGF-receptor complexes on cultured cells increases gradually from D = 2.8 X 10(-10) cm2/sec at 5 degrees C to 8.5 X 10(-10) cm2/sec at 37 degrees C. In the same range of temperature the rotational correlation times change from 25 to 50 microseconds to approximately 350 microseconds. Hence, at 4 degrees C, the occupied EGF receptors translate and rotate rapidly in the plane of the membrane. At 37 degrees C, EGF receptors form microclusters composed of 10 to 50 molecules. Moreover, it is concluded that both at 4 degrees C and 37 degrees C lateral diffusion of the occupied receptors is not the rate determining step for either receptor clustering or internalization. EGF receptor is a 150,000 to 170,000 dalton glycoprotein. The receptor is in close proximity to an EGF-sensitive, cAMP-independent, tyrosine-specific protein kinase which also phosphorylates the receptor molecules itself. The EGF sensitive kinase is similar to the kinase activity which is associated with certain RNA tumor viruses. The fact that the non-mitogenic cyanogen-bromide cleaved EGF is as potent as native EGF in stimulating phosphorylation suggests that EGF-induced, protein phosphorylation is a necessary but insufficient signal for the induction of DNA synthesis by EGF. EGF receptor serves also as the binding site for Transforming Growth Factors (TGF) which compete with EGF and induce anchorage-independent growth of normal cells in soft agar. Tumor promoters such as phorbol ester effect the binding of EGF to its membrane receptors and its ability to stimulate DNA synthesis. EGF itself has also some tumor promoting activity. Hence, the membrane receptor for EGF seems to participate in the regulation of normal and neoplastic growth. Monoclonal antibodies against EGF receptor (IgM) induce various early and delayed effects of EGF, while their monovalent Fab' fragments are devoid of biological activity. These observations support the notions that EGF receptor rather than EGF itself is the active moiety and that the role of the hormone is to perturb the receptor in the appropriate way, probably by inducing the microaggregation of EGF receptors.

Genetics of receptors for bioactive polypeptides: expression of the human EGF receptor gene and internalization and processing of the receptor-bound EGF in human-mouse cell hybrids

We previously postulated that the structural gene for epidermal growth factor (EGF) receptor is located on human chromosome 7 (1,2). In this study, EGF receptor and certain postreceptor functions were further analyzed in a unique cell hybrid line, C2B5, that retains only one human chromosome of an X;7 translocation besides a nearly complete mouse parental genome. Kinetics and Scatchard analysis of [125I]EGF binding to the C2B5 hybrid cells indicated that they carry a single class of EGF receptors with a dissociation constant of 4 x 10(-10) M. The receptors expressed in the hybrids are proven to be immunologically of human nature. The human EGF receptors now embedded in essentially mouse plasma membrane are subject to "down regulation" mediated by the ligand EGF. Analysis of the cell-bound EGF indicated that internalization and processing take place in the human-mouse cell hybrids. The degradation of EGF appears to be through a lysosomal pathway since it was substantially delayed or inhibited by lysosomotropic agents.

Properties of receptors for epidermal growth factor in detergent solution: evidence for heterogeneous aggregated states

Between 60% and 100% of epidermal growth factor (EGF) binding activity was recovered from membranes of the A431 human epidermoid carcinoma cell line treated with solutions containing the nonionic detergent Triton X-100. Approximately half of the recovered binding activity was sedimented at low centrifugal force and hence was operationally insoluble in nonionic detergent solution. Receptors in both the detergent-soluble and -insoluble fractions displayed similar affinities for 125I-EGF, and the values were in good agreement with those obtained for receptors in untreated membranes. The receptors in both fractions also formed identical direct linkage complexes with 125I-EGF in similar yield, providing no evidence for partitioning of different molecular species of EGF receptors in the detergent-soluble membrane fraction of Sepharose 6-B revealed heterogeneity of 125I-EGF binding activity; the smallest and most monodisperse peak of activity resolved by this technique was eluted at a Strokes radius of 95 A. Operationally soluble 125I-EGF binding activity also behaved heterogeneously during velocity sedimentation; more than half the activity sedimented more rapidly than the apparently monidisperse, 7S form. An average of less than half the nonionic detergent-solubilized activity recovered from 10 independent membrane preparations behaved as an apparently monodisperse entity. Since a maximum of 60% of 125I-EGF binding activity was operationally soluble, less than 25% of the total EGF binding activity was recovered in an apparently monodisperse form. The remaining 75% of the EGF receptors displayed a marked tendency to exist as aggregates in nonionic detergent solutions.

Controlled proteolysis of EGF receptors: evidence for transmembrane distribution of the EGF binding and phosphate acceptor sites

A small quantity of the 125I-EGF (epidermal growth factor) bound specifically to EGF receptors on the human epidermoid carcinoma cell line A431 associates covalently. The direct linkage complex formed migrates during gel electrophoresis as a single diffuse band of MW = 160,000-170,000. In contrast, direct linkage complexes of 160,000, 145,000, and 115,000 daltons are formed when EGF is incubated with membranes isolated from these cells; these arise from EGF receptor modification during membrane isolation. None of these modifications affected the affinity of the EGF binding site for 125I-EGF. The electrophoretic mobilities of the MW = 160,000 and 145,000 direct linkage complexes were similar to those of the major 32Pi-labeled products of the EGF-stimulated phosphorylation reaction described by Carpenter et al [Nature 276:409-410, 1978], indicating that proteolytic fragments of EGF receptors are the major phosphate acceptors in this reaction. EGF receptors on intact A431 cells accepted phosphate effectively from gamma-32Pi-ATP only when the cells were permeabilized with lysolecithin. This shows that the EGF binding and phosphate acceptor sites lie on opposing faces of the membrane. When the 145,000 dalton form of receptor is labeled with EGF or 32Pi and the labeled peptides subjects to tryptic hydrolysis under identical conditions, all phosphate is lost from high molecular weight products under conditions where the EGF-receptor covalent complex is converted largely to a 115,000 dalton form. This suggests that the phosphate acceptor site lies on the cytoplasmic side of the membrane on a region of receptor extending 30,000 daltons from the 115,000 dalton fragment containing the EGF binding site.