Chapter 1: Literature review 1 A REVIEW OF THE EPH/EPHRIN RECEPTOR TYROSINE KINASES 1.1 Overview of the Eph/ephrin receptor-ligand family The Eph receptor tyrosine kinases and their ephrin ligands are cell surface molecules with a wide range of biological functions which influence cell behaviour during both embryogenesis and adult life (Poliakov et al., 2004). These functions include roles in:- (1) the direction of cell positioning and migration (Klein, 2004); (2) axon guidance during development (Flanagan, 2006); (3) control of tissue morphogenesis and patterning (Coulthard et al., 2002; Hinck, 2004; Tepass et al., 2002); (4) defining tissue boundary formation during somitogenesis (Sela-Donenfeld and Wilkinson, 2005); (5) the development of the vascular system (Zhang and Hughes, 2006); (6) neural plasticity (Murai and Pasquale, 2004; Yamaguchi and Pasquale, 2004); (7) tumour invasion and metastasis (Wimmer-Kleikamp and Lackmann, 2005); (8) immune function (Wu and Luo, 2005), haematopoiesis (Okubo et al., 2006) and blood clotting (Prevost et al., 2005); (9) an emerging role in the biology of stem cells (Sancho et al., 2003) and (10) tissue repair and maintenance (Hafner et al., 2004). The Eph/ephrin receptor-ligand family, (which will be subsequently referred to jointly as the Eph/ephrin proteins), principally modifies cytoskeletal organisation and cell adhesion (Wilkinson, 2000). Cytoskeletal modification regulates the dynamics of cellular protrusions, affects cell–cell repulsion and adhesion and attachment to the extracellular matrix, and influences cell migration (Boyd and Lackmann, 2001). The signalling cascades that Eph-ephrin interactions initiate ultimately converge on targets such as integrins and small Rho-family GTPases (Noren and Pasquale, 2004). This is in contrast to many receptor tyrosine kinases (RTKs), which were first identified as oncogenes, because they activate signalling pathways that target gene transcription, and regulate cell proliferation and/or differentiation (Schlessinger and Ullrich, 1992; Ullrich and Schlessinger, 1990). The Eph/ephrin proteins, which constitute the largest class of RTKs in the human genome, are conserved amongst vertebrates, insects, nematodes and sponges (Drescher, 2002; Gu and Gu, 2003; Tuzi and Gullick, 1994; Vidovic et al., 2006). The Eph/ephrin family of receptors are so designated because they are related to Eph (pronounced Queensland Institute of Medical Research 1 Chapter 1: Literature Review “eff”), the first receptor described, which was named for its expression in an erythropoietin producing human hepatocellular carcinoma cell line (Hirai et al., 1987). The ligands of the Eph receptors are known as ephrins (pronounced “effrins”), an abbreviation derived from Eph family receptor interacting proteins. There are ten EphA receptors, EphA1–EphA10 (pronounced “eff-A”), and six EphB receptors, EphB1– EphB4 and EphB6 in vertebrates (Eph Nomenclature Committee, 1997; Lemke, 1997), and an additional EphB receptor, EphB5, which exists in chickens (Wilkinson, 2001). The nematode Caenorhabiditis elegans possesses a single Eph receptor, Vab-1, and four ephrins, Vab-2/EFN1, EFN2–4 (George et al., 1998; Wang et al., 1999). The fruitfly Drosophila melanogaster has a single Eph receptor, Dek, and a single transmembrane ephrin, dephrin (Bossing and Brand, 2002; Scully et al., 1999). Figure 1.1 Eph receptor and ephrin ligand binding preferences. (Modified from Wilkinson 2000). The initial distinction between EphA and EphB receptors was based on sequence differences within the extracellular ligand binding domain, but also corresponds to the binding preferences for the six glycosylphosphatidylinositol (GPI)-linked ephrin-A ligands and the three transmembrane ephrin-B ligands respectively (Eph Nomenclature Committee, 1997; Gale and Yancopoulos, 1997; Himanen and Nikolov, 2003a; Lemke, 1997) (Figure 1.1). 2 Queensland Institute of Medical Research Chapter 1: Literature review The Eph/ephrin receptor-ligand interactions are promiscuous within each A or B class, with variations in binding affinities (Flanagan and Vanderhaeghen, 1998), although EphB4 only binds ephrin-B2 (Blits-Huizinga et al., 2004). There are also exceptions in the binding preferences between A and B class, as EphA4 binds to ephrin-B ligands (ephrin-B2–ephrin-B3) (Blits-Huizinga et al., 2004), and EphB2 binds to ephrin-A5 (Himanen et al., 2004). 1.2 The structural basis for Eph/ephrin receptor-ligand interaction The Eph RTKs are type-I transmembrane proteins (Hubbard, 1999; Ullrich and Schlessinger, 1990). Their extracellular portion includes an 180-amino acid N-terminal “globular” ligand (ephrin) binding domain (Labrador et al., 1997), a cysteine rich region containing an epidermal growth factor (EGF)-like motif and two fibronectin type III repeats (Himanen and Nikolov, 2003b). This is separated by a transmembrane segment from the cytoplasmic portion of the receptor, which includes a juxtamembrane segment, tyrosine kinase domain, a sterile alpha motif (SAM) domain (Schultz et al., 1997) and a PDZ (post-synaptic density protein-95, Drosophila disc large tumour suppressor (Dlg), zona occludens-1) (Fanning and Anderson, 1999; Himanen and Nikolov, 2003a) (Figure 1.2). The ephrin receptor-binding domain also contains a globular N-terminal binding, or receptor-binding domain, which is separated from the membrane by a linker of approximately 40 amino acids (Himanen and Nikolov, 2003a; Himanen and Nikolov, 2003b). The ephrin-A ligands are attached via a GPI linkage to the plasma membrane (Himanen and Nikolov, 2003a; Himanen and Nikolov, 2003b). The ephrin-B ligands have a short highly conserved cytoplasmic domain consisting of 80 amino acids, which includes a C-terminal PDZ-binding motif (Gauthier and Robbins, 2003; Himanen and Nikolov, 2003a; Himanen and Nikolov, 2003b; Lin et al., 1999). The extracellular domain of the EphB2 receptor has a compact globular structure with a Q-sandwich “jellyroll” folding topology, which comprises two anti-parallel Q-sheets, a concave sheet composed of six strands and a convex sheet composed of seven strands, which together form a compact Q-sandwich that has an extensive hydrophobic core (Himanen et al., 1998). The Q-strands are connected by loops of varying length; in particular a long well-ordered loop (H–I) and two partially disordered loops which play a central role in ligand recognition and binding (Himanen et al., 1998). This Q-jellyroll Queensland Institute of Medical Research 3 Chapter 1: Literature Review folding topology shares no precise sequence or structural homology to other known proteins, although a similar folding topology is found in some carbohydrate binding proteins (lectins). Figure 1.2 The structure of the Eph receptors and their ephrin ligands. The EphA and EphB receptors have a conserved domain structure. The ephrin-A ligands are attached to the cell membrane by a GPI anchor. The ephrin-B ligands are transmembrane proteins. (Modified from Murai and Pasquale, 2003) The only sequence feature that is conserved within but differs between the EphA and EphB receptor classes is the length of the H–I loop, and consequently this loop is referred to as the class-specificity loop. The H–I loop consists of 17 amino acid residues in EphA receptors and 13 amino acid residues in EphB receptors (Himanen et al., 1998). 4 Queensland Institute of Medical Research Chapter 1: Literature review In mutagenesis experiments, a chimaeric EphB2 receptor constructed with an EphA3 H–I loop recognised both ephrin-A and ephrin-B ligands (Himanen et al., 1998). SAM is a protein interaction domain of approximately 70 amino acids, which was identified originally in yeast sexual differentiation and Drosophila proteins. Subsequently, SAM domains have been identified in over 100 proteins of fungi, protozoa and animals, in which they play a role in developmental regulation (Schultz et al., 1997). All the Eph receptors contain a highly conserved C-terminus SAM domain with approximately 50% identity (Stapleton et al., 1999). The X-ray crystal structure of the SAM domain of EphA4 has been resolved and consists of four short -helices and a longer C-terminal -helix in a “lobster claw” configuration, which form homo-dimers within both a crystal environment and in solution (Stapleton et al., 1999). The ephrins also contain a conserved extracellular receptor binding domain, which has novel topology for a signalling protein, shared only with plant phytocyanins and nodulins. X-ray crystallographic studies of ephrin-B2 indicate that the ephrin receptor binding domain has a globular Q-barrel structure with a Greek key folding topology, and crystals of uncomplexed ephrin-B2 form homo-dimers by burying the hydrophobic regions around the G–H loop (Toth et al., 2001). The G–H loop is involved in receptor binding, so it is likely that there is significant rearrangement of the ephrin homo-dimers upon Eph receptor interaction (Himanen et al., 2001). In contrast, ephrin-B1 is a monomer and its ectodomain has a distinct G–H loop conformation, which is not involved in homo-dimerisation as has been reported for ephrin-B2. Ephrin-B1 is glycosylated at Asn 139 in contrast to ephrin-B2 which is glycosylated at Asn 39, which may account for the lower binding affinity of ephrin-B1 to EphB2 (Nikolov et al., 2005). The entire cytoplasmic domain of ephrin-B2 has been studied using nuclear magnetic resonance (NMR) spectroscopy, revealing that the 48 N-terminal residues of the cytoplasmic domain are unstructured and prone to aggregation,