Epidermal Growth Factor and Receptor
Created by Monica Li
Epidermal growth factor (EGF) and epidermal growth factor receptor (EGFR) from Homo sapiens constitute a receptor-peptide complex involved in cell proliferation (1). Overexpression, deregulation, and mutation of EGF and EGFR are observed in many types of epithelial cancers, including breast, lung, and prostate cancers (2). As such, the structure and function of EGF and EGFR continue to be the subject of much study. Specifically, a 2:2 complex of EGF bound to the extracellular domains of EGFR (pdb ID = 1IVO) has been crystallized. The crystal structure was determined at a resolution of 3.3 Å using multiwavelength anomalous dispersion (MAD) (1). The EGF:EGFR homodimer has a molecular weight of 149,788.58 Da and an isoelectric point (pI) of 6.20.
In the absence of EGF, EGFR exists as an inactive monomer. When EGF binds to EGFR, the receptor-ligand complex dimerizes into its active form (1). Dimerization activates the tyrosine kinase activity of the intracellular domains of EGFR, thereby initiating receptor trafficking and cell signaling pathways that lead to cell survival and proliferation (3). Because this cascade of cellular activity is initiated by ligand binding and receptor dimerization, it is important to study the structural complementarity involved in the EGF:EGFR and the EGFR:EGFR interactions. These interactions are also clinically significant, as mutations and upregulation of EGFR are linked to uncontrolled cell proliferation, inhibition of apoptosis, angiogenesis, and tumor cell metastasis (4).
EGF is a small polypeptide with 53 amino acid residues and a molecular weight of 6,222.01 Da. The polypeptide is composed of 75% hydrophilic residues, which suits its function as a water-soluble, circulating cell signaling molecule (5). EGF can be divided into the A (residues 6-19), B (residues 20-31), and C (residues 33-42) loops. The A loop contains a short alpha helix-like segment, while the B and C loops each contain a short antiparallel beta sheet. The two strands of the B loop beta sheet are linked by a beta bend. The polypeptide contains six cysteine residues, each of which participates in an intrachain disulfide bond. These disulfide linkages define the three loops of the protein (1).
The extracellular domains (domains I-IV) of EGFR comprise a polypeptide chain with a length of 622 residues and a molecular weight of 68,699.30 Da. Domain I (residues 1-165) and domain III (310-481) have similar secondary structures (6). Each of these domains contains a short alpha-helical segment and two regions of parallel beta sheets, which form a right-handed beta barrel (1). Domain II (residues 166-309) and domain IV (residues 482-621) are cysteine-rich domains (6). The secondary structure of domain II is largely composed of short, antiparallel beta sheets and random coils. Dimerization of two EGFR molecules occurs via surface interactions involving domain II of each molecule, while the binding pocket for EGF is composed of portions of domains I and III (1). Domain IV, which contains less orderly secondary structure than domains I-III, serves to connect the extracellular domains to the transmembrane domain. The cytoplasmic side of the transmembrane domain is in turn connected to the intracellular tyrosine kinase domains (6).
The extracellular domains of EGFR contain 12 Asn-linked glycosylation sites (6). All of these sites are glycosylated in vivo (7). However, the protein was partially deglycosylated prior to crystallization; therefore, the crystal structure contains nine molecules of N-acetylglucosamine covalently linked to nine Asn residues (1).
The tertiary structures of EGF and the extracellular domains of EGFR are directly related to their functions. Certain aspects of the conformations of the two proteins allow for receptor-ligand interactions. At site 1, the B loop of EGF binds to domain I of EGFR via hydrophobic interactions. In particular, Ile-23 of EGF rests in a hydrophobic binding pocket formed by Leu-14, Tyr-45, and Leu-69 of EGFR. Mutagenesis studies show that replacing Ile-23 with any other hydrophobic residue significantly reduces binding activity, suggesting that the shape of the isoleucine side chain is critical for the receptor-ligand interaction. Also involved at site 1 is a hydrogen bond between the side chains of Asn-32 of EGF and Gln-16 of EGFR. At site 2, the A-loop of EGF binds with domain III of EGFR. Critical interactions in this region include the aromatic stacking of Tyr-13 of EGF and Phe-357 of EGFR and the ionic attraction between Arg-41 of EGF and Asp-355 of EGFR. At site 3, the C-terminal region of EGF binds with domain III of EGFR. Here, Leu-47 of EGF participates in hydrophobic interactions with Leu-382, Phe-412, and Ile-438 of EGFR (1). These noncovalent interactions all demonstrate how structural complementarity between EGF and EGFR facilitates their binding.
Physiologically, the binding of EGF to EGFR triggers the dimerization of the EGF:EGFR complex. The portion of EGFR involved in dimerization is a protruding region of domain II called the dimerization arm. Hydrophobic interactions between Tyr-246 and Tyr-251 of one EGFR molecule and Phe-230, Phe-263, and Ala-265 of another EGFR molecule hold the two receptors together. These five residues are all conserved or substituted with similar amino acids across diverse phylogenetic classes, indicating that they are vital to the receptor-receptor interaction. Additionally, a hydrogen bond between Tyr-251 of one EGFR molecule and Arg-285 of the other EGFR molecule stabilizes the dimer. Finally, mutagenesis studies show that the interface of domains II and III of a single EGFR molecule is important for dimerization: when the ionic interaction of Glu-293 and Arg-405 is destroyed, EGFR signaling activity is eliminated. It is thought that dimerization brings the tyrosine kinase domains of EGFR in close proximity with each other. Autophosphorylation then takes place, initiating complex cell signaling pathways (1).
Some oncogenic mutant versions of EGFR are constitutively active and spontaneously dimerize in the absence of EGF binding (1). In addition, overexpression or deregulation of EGFR has been detected in a diverse array of epithelial cancers. Therefore, several anti-cancer drugs are targeted at EGFR. For example, cetuximab is a monoclonal antibody indicated for the treatment of advanced colorectal cancer. The complex of the antigen-binding fragment of cetuximab and the extracellular domains of EGFR has been crystallized (pdb ID = 1YY9). The light chain of cetuximab binds to domain III of EGFR via hydrophobic interactions and a hydrogen bond between Gln-27 of cetuximab and Asn-473 of EGFR. In particular, the phenyl ring of Tyr-102 of cetuximab rests in the hydrophobic pocket of EGFR that would normally be occupied by Leu-47 of EGF. By blocking one of the EGF binding sites in domain III, cetuximab inhibits the dimerization of EGFR and the tyrosine kinase cascade that it triggers (2).
EGFR is capable of binding with several other EGF-like growth factors (2). One of these molecules is transforming growth factor alpha (TGFα), a 50-residue polypeptide. A crystal structure of domains I-III of EGFR complexed with TGFα (pdb ID = 1MOX) shows that TGFα, like EGF, binds to the receptor at the beta sheets of domains I and III. Once bound, the TGFα:EGFR complex dimerizes. Two dimer structures are observed: one in which two EGFR molecules interact via cysteine-rich domain II (analogous to the EGF:EGFR homodimer), and another in which the interaction takes place via domains I and III (8).
Because of its vital role in cell signaling pathways, the sequence and structure of epidermal growth factor are highly conserved across many animal species. A BLAST search was conducted to find proteins with primary structures similar to those of EGF and EGFR. It was found that murine epidermal growth factor (pdb ID = 1EGF) from Mus musculus, a 53-residue polypeptide, has 70% sequence identity with human EGF (E = 6 x 10-29). Out of the 16 amino acid differences, seven are substitutions with homologous amino acids, such as Asp-40/Glu-40 and Ile-35/Val-35 in murine EGF and human EGF, respectively. All six cysteine residues are conserved between the two proteins, and they form the same three disulfide bonds and define the same three loops. The secondary structure of murine EGF includes a single antiparallel beta sheet comprising three strands. One of these strands is a two-residue segment very close to the N-terminus of the polypeptide; this feature is absent from human EGF. Due to the alignment of this strand with the rest of the beta sheet, murine EGF has a more compact shape than human EGF (9).
Similarly, the primary and tertiary structures of epidermal growth factor receptors exhibit a large degree of conservation. A BLAST search revealed that the extracellular domains of Drosophila melanogaster EGFR (pdb ID = 3LTF) have 39% sequence identity with the corresponding domains of human EGFR, with an E-value of 3 x 10-147 (10). Additionally, a Dali server query showed that the receptors from the two organisms also have similar tertiary structures (Z = 36.2) (11). Like human EGFR, Drosophila EGFR exhibits beta barrel secondary structure in the ligand-binding domains I and III and contains a dimerization arm in domain II. However, when Drosophila EGFR binds spitz, an EGF-like growth factor, the complex aggregates into an asymmetrical dimer with two structurally inequivalent ligand binding sites (10). By comparison, the human EGF:EGFR dimer is symmetrical and contains two identical EGF binding sites (1). The asymmetry of the Drosophila spitz:EGFR dimer and the inequivalence of the binding sites suggest that cooperativity is involved in spitz binding and EGFR dimerization in Drosophila (10).