Mechanisms of activation of the guanine nucleotide exchange factor C3G

Resumo

C3G (also known as RapGEF1) is a ubiquitously expressed guanine nucleotide exchange factor (GEF) that activates mainly the small GTPase Rap1 to promote integrin-mediated cell adhesion. C3G also regulates cell migration, actin remodeling, proliferation, apoptosis, differentiation, and exocytosis. C3G is a modular protein with three structurally and functionally distinct regions. The N-terminal domain (NTD) interacts with E-cadherin and participates in the self-regulation of C3G. The central or SH3b region is mostly flexible and is involved in protein-protein interactions. This region contains five proline-rich motifs (PRMs), P0 to P4, which are binding sites for SH3 domains. The Crk and CrkL adaptor proteins bind to four of these sites (P1, P2, P3 and P4). The C-terminal part contains a REM and a Cdc25H domain that form the catalytic region. The GEF activity of C3G is regulated by two intramolecular interactions. First, binding of the NTD to the REM domain contributes positively to the activity of C3G. Secondly, the C-terminal segment of the SH3b downstream the P3 is an autoinhibitory region (AIR) that binds to the Cdc25HD and blocks the GEF activity. C3G is physiologically activated in response to stimuli that operate through the activation of tyrosine kinases. These signals induce the recruitment of C3G to signaling sites at the membrane, which is mediated by Crk proteins. Activation of the GEF activity of C3G involves tyrosine phosphorylation by Src family and other kinases, and the interaction with Crk proteins. Despite the general processes involved in C3G activation were known, the detailed mechanisms were poorly understood. In this Thesis, we have combined biochemical, biophysical and cell biology approaches to gain understanding of the structural and mechanistic basis of the autoinhibition and physiological activation mechanism of C3G activation within the C3G-Rap1 pathway. We have demonstrated that CrkL binds differentially to the PRMs P1, P2, P3 and P4 in full-length C3G. In resting conditions, sites P1 and P2 are constitutively fully accessible. Exposure of the P3 site is linked to the activation state of C3G. The P4 site is partially accessible independently of the activation state of C3G. The sites P1 and P2 do not participate in the direct activation of C3G. Instead, these sites participate in the constitutive interaction with Crk proteins in HEK293T and Jurkat cell lines, and are required for the recruitment of C3G to the plasma membrane upon TCR stimulation of Jurkat cells. Binding of CrkL to the P3 and P4 is required and sufficient for the direct activation of C3G by CrkL through the release of the autoinhibitory interaction. Binding of CrkL to the P3 is the main activation event. Collectively, P1 and P2 are recruitment sites and P3 and P4 are activation sites. Tyrosine phosphorylation of C3G alone causes minimal activation, yet it is essential for the Crk-mediated stimulation. Phosphorylation by Src sensitizes C3G, which is activated at lower concentrations of CrkL and to higher activity levels than unphosphorylated C3G. We have mapped the main Src phosphorylation sites in C3G at Y329, Y504, Y579 and Y590. Despite CrkL interacts with the P3 and P4 primarily by the SH3N domain, the SH2 and SH3C domains of CrkL also contribute to the activation of C3G. In particular, the SH2 domain plays a key role in the activation of phosphorylated-C3G through a non-canonical low affinity interaction with phospho-Y590, which apparently stabilizes the interaction of CrkL with the activation sites. We have also shown that CrkL induces a stronger activation of C3G than CrkII, and these differences are apparently due to the different inter-domain arrangements of Crk proteins. We have also produced an experimentally validated structural model of autoinhibited C3G, which revealed additional contacts of the AIR with the REM domain. Based on the results, propose a multi-step cycle for the physiological activation and de-activation of C3G. Finally, we have applied the structural and functional data to identify three cancer somatic missense mutations that cause constitutive activation of C3G: Y570N and Y590N were found in non-Hodgkin lymphoma patients, and Y579C was described in a thyroid carcinoma. These mutations expand the repertoire of acquired alterations that deregulate C3G-Rap1 in cancers. In summary, the results of this Thesis contribute to understand the mechanisms of C3G-Rap1 signaling in healthy tissues and in diseases.