Role of extracellular subdomains of p185^c-neu and the epidermal growth factor receptor in ligand-independent association and transactivation

Abstract

We investigated the assembly and activation of the epidermal growth factor receptor (EGFR)–p185^c-neu heterodimer by using a sequential immunoprecipitation methodology. Using this approach we detected heterodimers and also higher-ordered oligomeric complexes. Phosphorylated EGFR–p185^c-neu heterodimeric forms were detected in the absence of EGF, but the species became highly phosphorylated after EGF stimulation. To evaluate heterodimer formation and additional transactivation by EGF, we investigated the roles of the four extracellular subdomains of p185^c-neu and the EGFR. Subdomains I–IV of the EGFR dimerized with subdomains I–IV of p185^c-neu, respectively, in a parallel manner. In addition, subdomains I–IV of the EGFR also associated with p185^c-neu subdomains III, IV, I, and II, respectively. A lack of one of the p185^c-neu cysteine-rich domains (subdomains II or IV) resulted in a loss of EGF-induced transactivation. These data suggest that two cysteine-rich domains play defining roles in ligand-dependent transactivation and that both of these cysteine-rich extracellular subdomains as well as non-cysteine-rich extracellular subdomains are involved in ligand-independent interactions with the EGFR. Our studies provide biochemical evidence of the role of the cysteine-rich domains of p185^c-neu in assembly and transactivation of erbB complexes and also indicate that these subdomains might be useful clinical targets.

The erbB receptor family of tyrosine kinases includes four members: erbB-1 [the epidermal growth factor receptor (EGFR)], erbB-2 (the human homologue of rat neu, p185^c-neu), erbB-3, and erbB-4 (1–7). These receptors play important roles in differentiation, proliferation, development, and transformation. Diversity may result from dimerization that can occur between these receptor members (8–16). In addition, the erbB family receptors seem able to participate in diverse signaling pathways with other receptor families (17). Homo- or heterodimerization of erbB members is needed for activation of the receptor complex and subsequent signal transduction (16, 18–24). It has been proposed that once EGFR dimerization is induced by ligand, the tyrosine kinase subdomain is activated and tyrosine residues at the C terminus are autophosphorylated, allowing them to serve as docking sites for other diverse signaling molecules (16, 21, 25). Among erbB receptors, p185^erbB2/c-neu is the preferred heterodimerization partner for other erbB receptors, although the basis of this preference is not understood (16, 18, 19, 23).

Ligands such as EGF stimulate p185^c-neu–EGFR heterodimers, activating the kinase complex and leading in some cases to transformation (8, 18, 26). It is well known that overexpression of either erbB2 or EGFR protein is frequently observed in human cancer (27–31), and in this regard it is compelling that overexpression of both (erbB2 and EGFR) receptors correlates with the poorest prognosis in breast carcinomas (30, 32). We have made efforts to clarify the mechanisms by which assembled erbB receptors are activated and become able to mediate transformation (8, 18, 26).

EGF clearly binds the p185^erbB2/c-neu–EGFR heterodimer (18, 22, 33). Lemmon et al. (34) suggested that two EGF molecules associate with the EGFR homodimer. In this context, the homodimer of the EGFR has two affinity states for EGF (low and high), but three affinity states (a very high state in addition to the former two) for EGF are observed by coexpression of p185^c-neu and the EGFR. Indeed, the very high-affinity binding state is only seen with heterodimeric forms that are created by associations of the EGFR and p185^c-neu (18). However, Ferguson et al. (24), using plasmon resonance assays in vitro, showed that a truncated extracellular domain fragment of the EGFR does not form heterodimeric species with ectodomain fragments of erbB2 protein after EGF stimulation. We reported the possible existence of heterooligomeric forms other than dimers based on modeling and structural considerations of kinase assemblies (35, 36). In this study we attempted to detect heterooligomers, including heterodimers, to clarify which forms are affected by EGF.

p185^erbB2/c-neu has four extracellular subdomains including two cysteine-rich domains (subdomains II and IV). The functional significance of the two cysteine-rich domains is unclear (2, 4, 7). Previously we described a p185^c-neu form lacking most of the cytoplasmic domain (N691 stop) that exerted suppressive effects on the function of the EGFR (37). We also found that the extracellular subdomains I–II and IV of p185^c-neu were critical for dimerization with the EGFR and that subdomain IV with the transmembrane domain could suppress EGFR function through heterodimerization (36). This set of studies provided a compelling reason to continue the analysis of subdomain contributions to EGFR–p185^c-neu activation to elucidate common and general principles involved in receptor activation that might be explained by heterooligomerization. Our argument, supported by findings of other laboratories, has been that heterooligomerization is a common general mechanism to create signaling diversification of monomorphic receptors (18). However, the roles of the extracellular subdomains of the EGFR and p185^c-neu in terms of (i) heterodimerization or receptor assembly and (ii) transactivation have remained undefined.

In this study we identify previously undetected heteromeric forms between the EGFR and p185^c-neu, show the effect of EGF on transactivation of the heterodimer, and examine the role of individual extracellular subdomains on dimerization and transactivation.

Materials and Methods

Expression Vectors. A schematic representation of the expression vectors used in this study is shown in Fig. 1. These expression vectors were used as described (36). pNeu encodes full-length protooncogenic p185^c-neu; pNex3 and pNex4 encode extracellular subdomain deletion mutant forms of p185^c-neu containing only extracellular subdomain III or IV, respectively; and pMVEGFR encodes full-length EGFR.

Fig. 1.

Schematic representation of the expression vectors. Numbers, amino acid positions from the first Met at the N terminus of p185^c-neu or the EGFR; SP, signal peptide from wild-type EGFR; SPκ, signal peptide from murine Ig κ-chain leader sequence; I–IV, subdomains I–IV, respectively; TM, transmembrane domain; TK, tyrosine kinase domain; Myc, Myc epitope; His, polyhistidine tag; VSVG, vesicular stomatitis virus glycoprotein tag; HA, hemagglutinin tag; PDGFR, platelet-derived growth factor receptor; broken line, deleted region.

pNex encodes p185^c-neu lacking most of the intracellular domain that is identical to N691 (37) but also contains a c-Myc epitope tag and a polyhistidine tag at the C terminus. p2-4Neu, p3-4Neu, p4Neu, and pΔ4Neu lack extracellular subdomains I, I–II, I–III, and IV, respectively. These constructs were subcloned into pSecTagB (Invitrogen). We also developed mutant p185^c-neu forms that contain a single extracellular subdomain with the transmembrane domain subcloned into pSecTagB (pNex1 and pNex2) and mutant EGFR forms that contain a single EGFR extracellular subdomain subcloned into pDisplay (Invitrogen) containing the transmembrane domain of platelet-derived growth factor receptor (pE1, pE2, pE3, and pE4) as indicated in Fig. 1.

Cell Lines. Cos7 and NR6 cells were used in this study as described (36).

Transfection. Cos7 cells (1 × 10⁶ per 10-cm dish) were transfected with the various kinds of vectors by using the N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate (DOTAP) liposomal transfection reagent according to manufacturer protocol (Roche Molecular Biochemicals).

Heterooligomerization Studies. Cos7 cells were transfected with pMVEGFR and an equal amount of pNeu. Forty hours later, the medium was replaced with serum-free medium and incubated for an additional 24 h. Cells were stimulated with 50 ng/ml EGF (GIBCO/BRL) at 37°C for 5 min or left unstimulated and then treated by bis(sulfosuccinimidyl) suberate (BS³) cross-linker (Pierce) at 4°C for 30 min as described (38). After stopping the cross-linking reaction with 10 mM Tris·HCl (pH 7.5)/0.1 M glycine in saline, cells were lysed in TNE buffer containing 10 mM Tris·HCl (pH 7.4), 1% Nonidet P-40, 0.1% deoxycholic acid, 0.1% SDS, 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, and 1 μg/ml aprotinin. Before immunoprecipitation, anti-His antibody (Santa Cruz Biotechnology) was conjugated with protein G by using dimethyl pimelimidate (DMP, Pierce). Equal amounts of each cell lysate were immunoprecipitated with anti-His antibody conjugated with protein G-Sepharose. Immunoprecipitated samples were eluted with 0.1 M glycine (pH 2.5) neutralized with a 1:19 volume of 1.5 M Tris (pH 8.8) and then dialyzed with PBS at 4°C overnight. Dialyzed samples were rotated with protein G before the second immunoprecipitation to remove the first antibody and then immunoprecipitated with anti-EGFR antibody 1005 (Santa Cruz Biotechnology). Immunoprecipitated samples were subjected to SDS/PAGE analysis and transferred to nitrocellulose membranes. We used cross-linked α₂-macroglobulin treated with bis(sulfosuccinimidyl) suberate as a molecular mass marker in addition to protein molecular mass markers (Amersham Pharmacia). Immunoblotting was performed with anti-c-myc antibody (Roche Molecular Biochemicals). After stripping, the membranes were reprobed with antiphosphotyrosine antibody PY99 (Santa Cruz Biotechnology). The ratio of phosphorylated heterodimer to total heterodimer was approximated by laser scanning microdensitometry analysis by using NIH image software.

p185^c-neu Subdomains Involved in Transactivation. Cos7 cells were transiently transfected with the EGFR construct and each mutant p185^c-neu construct. They were stimulated with 50 ng/ml of EGF or unstimulated and then lysed in either of the following lysis buffers: Brij buffer containing 1% Brij 97, 50 mM Tris·HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, and 1 μg/ml aprotinin or radioimmunoprecipitation assay buffer containing 10 mM sodium phosphate (pH 7.4), 1% Triton X-100, 1% deoxycholic acid, 0.1% SDS, 150 mM NaCl, 1 μg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 2 mM EDTA, 1 mM sodium orthovanadate, 10 mM NaF, 10 mM sodium pyrophosphate, and 10 mM iodoacetamide. After immunoprecipitation with anti-EGFR antibody, immunoprecipitated samples were subjected to SDS/PAGE analysis, and immunoblotting was performed with anti-vesicular stomatitis virus glycoprotein antibody (Roche Molecular Biochemicals), which recognizes the EGFR, or anti-c-myc antibody and then PY99.

Analysis of p185^c-neu Subdomain Contribution to Ligand-Independent Association. Mutant p185^c-neu forms that contain a single extracellular subdomain with the transmembrane domain and mutant EGFR forms that contain a single EGFR extracellular subdomain containing the transmembrane domain of platelet-derived growth factor receptor were analyzed for association.

Briefly, Cos7 cells transfected with these mutant forms were treated with 3,3′-dithiobis(sulfosuccinimidyl propionate) (DTSSP, Pierce) without EGF stimulation. Cells were lysed in radioimmunoprecipitation assay buffer as described above. Immunoprecipitation was performed with anti-His antibody, and immunoblotting was performed with anti-hemagglutinin antibody (Roche Molecular Biochemicals).

Results

p185^c-neu–EGFR Heterodimer in the Absence of EGF Is Activated by EGF. We have studied heterodimeric associations that occurred between p185^c-neu and the EGFR. Cos7 cells were transiently transfected with the EGFR and pNeu (Fig. 1) and then were used for the heterooligomeric studies. Because p185^c-neu and the EGFR have similar molecular masses (185 and 170 kDa each), it was difficult to distinguish the molecular masses of homodimers from heterodimeric complexes by SDS/PAGE analysis. To enrich for heteromeric complexes, cell lysates were first immunoprecipitated with anti-His antibodies to detect the p185^c-neu species and then eluted and subsequently immunoprecipitated with anti-EGFR antibody for detection of the EGFR associated with p185^c-neu.

We used a 3.7% polyacrylamide gel for detection of oligomeric complexes and used α₂-macroglobulin treated with bis(sulfosuccinimidyl) suberate as a molecular mass marker. Monomeric, dimeric, and tetrameric species of α₂-macroglobulin were detected as approximate molecular masses of 180, 360, and 725 kDa, respectively (Fig. 2, indicated as marker).

Fig. 2.

Detection and analysis of heterodimer formation between the EGFR and p185^c-neu by using sequential immunoprecipitation. Cos7 cells (1 × 10⁶ per 10-cm dish) transiently transfected with 3 μg of each indicated vector were stimulated with 50 ng/ml EGF (+) or unstimulated (–) followed by the cross-linker treatment as described in Materials and Methods. Cell lysates were immunoprecipitated (IP) with the antibodies as indicated. The immunoprecipitated sample was subjected to SDS/PAGE. Immunoblotting was performed with the antibodies as indicated. The membrane on the left was stripped and reprobed with PY99 (Right). E, pMVEGFR; N, pNeu; D, dimer; HO, higher-ordered heterooligomer complex.

Heterodimers could be detected by the sequential immunoprecipitation methodology (Fig. 2, lanes 1 and 2). Unexpectedly, a higher oligomeric complex containing both holoreceptor p185^c-neu and the EGFR forms could also be detected. Although these species are phosphorylated even in the absence of EGF, adding EGF seemed not to induce further phosphorylation of these complexes. Although the details of the organization of the higher oligomeric forms and their biological significance are uncertain, it is clear that they can be readily detected by using our approach.

We evaluated whether EGF could promote both heteromeric complex formation and further activation of the already partially activated and assembled heterodimer. The relative signal intensity of phosphorylated heterodimer after EGF stimulation was compared with that of untreated cells to address this question. Although EGF was not able to induce greater amounts of heterooligomeric assemblies in these studies, we were able to determine that EGF could activate the heterodimer. Heterodimers became more phosphorylated (220%) after EGF stimulation than the nonstimulated heterodimer (Fig. 2 Left and Right, lanes 1 and 2). On the other hand, the higher-ordered heterooligomer complexes were not further activated by EGF, suggesting that the biological behavior of higher-ordered heterooligomers might be different from that of heterodimer.

Although heterodimeric formation did not require EGF stimulation, EGF did elevate the phosphotyrosine content of the heterodimer. It seemed possible that EGF can activate the heterodimer that was formed in the absence of EGF. To investigate this possibility, heterodimer formation in the absence of EGF and transactivation of heterodimer was evaluated further by extracellular subdomain analysis.

Contribution of Extracellular Subdomains of p185^c-neu to Ligand-Dependent Transactivation. To clarify which subdomain contributes to this phenomenon, we analyzed heteromeric complex formation between the EGFR and different mutant p185^c-neu species lacking one or several extracellular subdomains. The effect of EGF on the activation of heterodimer was analyzed by using Cos7 cells transiently transfected with both EGFR and mutant p185^c-neu. Then EGF was added to some cells. Although the p2-4Neu protein that contained extracellular subdomains II–IV showed ligand-dependent transactivation of p185^c-neu (Fig. 3 Top, lanes 3 and 4), the p3-4Neu species containing subdomains III–IV did not show ligand-dependent transactivation (Fig. 3A, lanes 5 and 6). Another immunoblot analysis with radioimmunoprecipitation assay buffer revealed that the phosphotyrosine content of 2-4Neu increased to 177% of its prestimulation level after EGF ligand treatment (data not shown). These results indicate that subdomain II (distal cysteine-rich domain from the cell membrane) is needed to create a ligand-dependent transactivatable heteromeric receptor complexes.

Fig. 3.

Analysis of subdomain contribution to ligand-dependent transactivation of p185^c-neu. Cos7 cells were transiently transfected with 5 μg of the EGFR and equal amounts of each mutant p185^c-neu form. They were stimulated with 50 ng/ml EGF or left unstimulated. Cells were lysed in Brij buffer as described in Materials and Methods and immunoprecipitated (IP) with the indicated antibody followed by immunoblotting with the indicated antibody. Each blot was stripped and reprobed with the indicated antibody. Neu, pNeu; 2-4Neu, p2-4Neu; 3-4Neu, p3-4Neu; 4Neu, p4Neu; 6CNNeu, p6CNNeu; ΔNeu, pΔ4Neu; EGFR, pMVEGFR.

On the other hand, deletion of subdomain IV (pΔ4Neu) resulted in a species that could cause ligand-dependent transsuppression (Fig. 3A, lanes 11 and 12). Another immunoblot analysis also demonstrated that the phosphotyrosine content of pΔ4Neu was reduced to 18.9% of its unsuppressed control after EGF stimulation (data not shown). We therefore conclude that both subdomains II and IV, known as the cysteine-rich domains, are necessary for ligand-dependent transactivation.

Contribution of Extracellular Subdomains of p185^c-neu to Ligand-Independent Association. We also observed ligand-independent associations between the EGFR and various mutant p185^c-neu forms and therefore analyzed which extracellular subdomains were relevant to this type of association. Mutant p185^c-neu forms that contain the entire or single extracellular subdomains with a transmembrane domain and mutant EGFR forms that contain the single EGFR extracellular subdomain with the transmembrane domain of platelet-derived growth factor receptor were used (Fig. 1). The entire extracellular domain of p185^c-neu (Nex) was able to associate well with EGFR subdomains II and IV and slightly with subdomain I, but no association was observed with subdomain III (Fig. 4). To clarify the contribution of the extracellular subdomains of p185^c-neu and those of the EGFR to ligand-independent association, we investigated the interactions between single subdomains of p185^c-neu and those of the EGFR. Using these same methods, we further investigated whether subdomain III of the EGFR was also unable to associate with the extracellular subdomains of p185^c-neu. Associations between these mutant forms were analyzed in the absence of EGF. Both subdomains I and III of p185^c-neu were able to associate with subdomains I and III of the EGFR (Fig. 4 A and C). Subdomains II and IV, cysteine-rich domains of p185^c-neu, associated with subdomains II and IV (Fig. 4 B and D), cysteine-rich domains of the EGFR, in what seem to be either parallel or antiparallel interactions. Exceptionally faint associations were observed between subdomain IV of p185^c-neu and subdomain I of the EGFR. However, the associations were much less dominant than those between the non-cysteine-rich domains (Fig. 4A).

Fig. 4.

Analysis of ligand-independent subdomain interactions between p185^c-neu and the EGFR. Cos7 cells were transiently transfected with 5 μg of each indicated vector, and associations were analyzed without EGF stimulation. Samples immunoprecipitated with indicated antibodies (IP) or whole-cell lysates (WCL) were subjected to immunoblotting with the indicated antibodies (Probe). E1, pE1; E2, pE2; E3, pE3; E4, pE4; N1, pNex1; N2, pNex2; N3, pNex3; N4, pNex4; Nex, pNex.

Taken together, p185^c-neu and the EGFR form both active heterodimers and a newly described higher heterooligomeric species. Heterodimers are more activated after EGF stimulation, and the cysteine-rich domains of p185^c-neu are necessary for ligand-induced activation. Both cysteine-rich domains of p185^c-neu as well as non-cysteine-rich subdomains are involved in ligand-independent association.

Discussion

Our work has focused on issues related to assembly and subsequent formation of activated receptor complexes composed of erbB members. Little is known of how heterooligomers of the EGFR and p185^erbB2/c-neu are formed and activated (16, 39).

NIH 3T3 stably transfected cell lines expressing the EGFR alone at levels comparable to that of human tumor lines A431 and MDAMB468 (2–4 × 10⁶ receptors per cell) have been reported to develop a transformed phenotype in the presence of EGF (40). Transformation of NR6 cells was never seen when transfected with either p185^c-neu or the EGFR at lower levels (2.5 × 10⁵ receptors per cell) (8). However, coexpression of both p185^c-neu and the EGFR at moderate receptor numbers (1–2 × 10⁵ receptors per cell) renders these cells completely transformed (8). p185^erbB2/c-neu is known to be able to cooperate with other erbB family receptors during tumor development, and it is thought that tumor cells acquire more potent growth stimulation through the activation of additional intracellular pathways achieved by the heteromeric assemblies (16, 18).

The assembly process of oligomers is complex. Ferguson et al. (24) have suggested that the mechanisms for homo- and heterooligomerization of erbB receptors may be different. Our interest was to clarify how both receptors are activated through oligomerization and to provide insight into the steps and subdomains of the receptors that are needed to create an active kinase complex.

EGF induces homodimerization of the EGFR, activates the EGFR tyrosine kinase, and induces autophosphorylation of the receptor (22, 41). However, the behavior of the heterooligomer was quite different. Compared with the EGFR alone, EGFR–p185^c-neu assemblies possess unique properties in active signaling because heterodimer formation was observed even in the absence of EGF. Heterooligomer formation was also apparent without cross-linkers (Fig. 3A, lane 2). Despite having some activity in the absence of ligand, the heterodimer is activated further in an efficient manner after EGF stimulation. Unexpectedly, by using the sequential immunoprecipitation method, we could detect higher-ordered heterooligomers (tetrameric forms). Although the complex was phosphorylated in the absence of EGF, it was not activated further by EGF, suggesting that the biological behavior of the heterodimer and higher-ordered heterooligomers was different.

Coexpression of both p185^erbB2/c-neu and the EGFR results in cells that possess a higher proliferation index than cells in which the EGFR is expressed alone and seems to involve the additional activity of recruited signaling molecules (16, 22). Wada et al. (18) showed dimer and heteromer formation and dimerization between the EGFR and p185^c-neu in M1 cells by using cross-linking reagents. Ligand-dependent transphosphorylation of p185^c-neu induced by EGF was also demonstrated in this cell line (8). In this study we observed ligand-dependent transactivation of the heterodimer without “ligand-dependent” heterodimer formation in Cos7 cells by using the same cross-linker, bis(sulfosuccinimidyl) suberate. Our data indicate that ligand (EGF) is not always necessary for heterooligomerization between the EGFR and p185^c-neu. However, as also shown, EGF is able to bind to the heterodimer formed between the EGFR and p185^c-neu (16, 22, 33) and trigger assembly-related kinase activity.

When we consider these facts, we conclude that p185^c-neu may already possess a suitable structure for heterooligomerization with the EGFR even in the absence of EGF (42). Such a ligand-independent complex is already activated to some extent but can clearly be activated further. Once EGF binds to the EGFR–p185^c-neu heterodimer, the heteromeric complex becomes more activated.

The erbB family of receptors all have four extracellular subdomains including two cysteine-rich domains (2, 7). These cysteine-rich domains are characterized by cysteine knots and cysteine-bonded loops, all characterized by a distinct pattern of disulfide bonds (43, 44). Although the contribution of these subdomains to ligand-dependent transactivation has not been investigated, we have shown that both cysteine-rich domains are important for ligand-dependent activation. Because EGF binds to the EGFR, it may mean that both binding of EGF to the EGFR and some ligand-induced alteration of the cysteine-rich domains of p185^c-neu may be necessary for homodimerization.

On the other hand, ligand-independent activation also seems to be important in p185^c-neu–EGFR complex heterooligomer formation. Our data show that not only cysteine-rich domains but also non-cysteine-rich domains can participate in heterodimerization/oligomerization. Interactions between non-cysteine-rich subdomains as well as those of cysteine-rich domains seemed to play very important roles in ligand-independent interaction between p185^c-neu and the EGFR. It should be noted that a putative EGF-binding site of the EGFR has been localized to surfaces on subdomains I and III (7). Notably, the cysteine-rich domains of p185^c-neu prefer to associate with cysteine-rich domains of the EGFR and fail to associate, for the most part, with non-cysteine-rich domains.

Our results suggest that the cysteine-rich domains of p185^c-neu play an important role in ligand-dependent transactivation through associations with cysteine-rich subdomains of the EGFR, suggesting that surfaces of the cysteine-rich domain may become allosterically activated into a more interactive surface by stimulation with EGF. However, heterooligomer formation itself may be initiated by unique cysteine-rich domain conformations of p185^c-neu that do not require ligand-induced structural alterations (42). It is subsequent changes of these cysteine-rich domains that would be triggered by ligand binding to subdomains I and III that would account for EGF-enhanced activation of already partially active complexes. Therefore, heterooligomer formation and subsequent ligand activation are separately definable processes. We have reported that cysteine-rich domain IV of p185^c-neu suppressed EGF-induced transformation of cells expressing the EGFR and also defined the possibility of peptidomimetics derived from this domain to act as therapeutic agents (36, 45). Because this cysteine-rich domain associates with the EGFR, peptidomimetics derived from this domain might be useful against human cancers expressing both EGFR and p185^c-neu/erbB2 by inhibiting their ligand-independent association and -dependent activation that requires the interaction of their cysteine-rich domains.

In conclusion, p185^c-neu and the EGFR form both active heterodimers and higher-ordered heterooligomers even in the absence of ligand. Heterodimers are activated further after EGF stimulation. The cysteine-rich domains of p185^c-neu are necessary for ligand-induced activation, suggesting that although association of heteromers does not require ligand-induced changes, the ligand further alters the structure of the receptor to increase its kinase activity. These observations are important for understanding the activation mechanisms of EGFR–p185^c-neu heteromers, which lead to cellular transformation. Finally, although our work has dealt with erbB members, it is likely that the general principles of unique steps in creating and then activating homomeric and heteromeric assemblies of receptors will apply to other receptor families that diversify their signaling repertoire by forming complex assemblies.