br Experimental procedures br Acknowledgments br Introductio
Introduction Receptor tyrosine kinases (RTKs) control many fundamental cellular processes, such as cell proliferation, differentiation, migration, and metabolism (Lemmon and Schlessinger, 2010). RTK activity is normally tightly controlled, and dysregulation of RTK activity is associated with many human cancers and other pathologies. Ligand binding to the extracellular region of RTKs leads to autophosphorylation of their cytoplasmic kinase domains, creating docking sites for effectors of downstream signaling. The two major strategies for controlling unwanted RTK activity in human patients are inhibition by monoclonal Tempol (mAbs) directed against their extracellular regions or by small molecules targeting the kinase active site (Adams and Weiner, 2005, Gschwind et al., 2004). The discoidin domain receptors, DDR1 and DDR2, are RTKs that are activated by several types of triple-helical collagen, a major component of the animal extracellular matrix (Leitinger, 2011, Shrivastava et al., 1997, Vogel et al., 1997). The DDRs are widely expressed in mammalian tissues and have important roles in embryo development and human disease (Vogel et al., 2006). For example, DDR1 is essential for mammary gland development (Vogel et al., 2001), and DDR2 is essential for the growth of long bones (Labrador et al., 2001). DDR2 mutations in humans cause a rare, severe form of dwarfism (Ali et al., 2010, Bargal et al., 2009). The DDRs are also implicated in cancer, fibrotic diseases, atherosclerosis, and arthritis (Vogel et al., 2006). Mechanistically, the DDRs have several features that distinguish them from other RTKs. Compared with the rapid response of typical RTKs to their soluble ligands (e.g., growth factors), collagen-induced DDR autophosphorylation is slow and sustained (Shrivastava et al., 1997, Vogel et al., 1997). Furthermore, Src kinase plays an essential role in DDR activation (Ikeda et al., 2002). Both DDRs are composed of an N-terminal discoidin (DS) domain (Baumgartner et al., 1998), followed by a predicted DS-like domain (our unpublished results; Lemmon and Schlessinger, 2010), an extracellular juxtamembrane (JM) region, a transmembrane (TM) helix, a large cytosolic JM region, and a C-terminal tyrosine kinase domain. Collagen binds to the DS domain, and the structural determinants of the DDR-collagen interaction have been extensively studied (Carafoli et al., 2009, Ichikawa et al., 2007, Konitsiotis et al., 2008, Leitinger, 2003, Xu et al., 2011). The remainder of the extracellular region has not been characterized structurally or functionally. How collagen binding results in DDR activation is a major unresolved question. DDR1 can be activated by short collagen-like peptides, showing that DDR clustering by multivalent collagen assemblies (e.g., fibrils) is not essential for activation (Konitsiotis et al., 2008). The DDRs are constitutive dimers at the cell surface, and residues within the TM helix are required for signaling (Noordeen et al., 2006). In fact, a comprehensive analysis has shown that the DDRs have the highest propensity of TM helix self-interactions in the entire RTK superfamily (Finger et al., 2009). Therefore, the conformational changes resulting from collagen binding are likely to occur in the context of a stable DDR dimer. Our crystal structure of a DDR2 DS-collagen peptide complex (Carafoli et al., 2009) revealed a 1:1 complex and did not clarify how collagen binding affects the conformation of the DDR dimer. Here, we report the functional characterization of a set of inhibitory anti-DDR1 mAbs and the crystallization of the almost complete extracellular region of DDR1 bound to a mAb Fab fragment. The crystal structure led to the discovery of DDR1 residues that are required for signaling, even though cis-acting protein are not part of the known collagen-binding site. These results provide insight into the process of DDR1 activation.