Arrestins are multifunctional signaling proteins, important for the regulation of signal transduction and the trafficking of G protein-coupled receptors (GPCRs). Recently, GPCR-arrestin interactions have been proposed to be necessary for activation of G-protein-independent signaling pathways, one of which is the activation of mitogen activated protein kinases (MAPKs). To investigate potential arrestin-MAPK interactions, we have used a variety of molecular tools including the co-expression of the individual domains of arrestin with single components of the c-Raf1-MEK1-ERK2 signaling cascade. We found that non-visual arrestins bind all three kinases, assembling c-Raf1, MEK1, and ERK2 along their short axis, with each kinase directly interacting with both domains.
To further investigate the interactions between arrestins and MAPK, we used alanine-scanning mutagenesis of residues on the non-receptor-binding surface of arrestin that are conserved between arrestin-2 and arrestin-3. We found that the substitution of arginine 307 with an alanine significantly reduced arrestin-2 binding to c-Raf1, whereas the interactions of this mutant with active phosphorylated receptors and the downstream kinases MEK1 and ERK2 were not affected. In contrast to wild type arrestin-2, Arg307Ala mutant failed to rescue arrestin-dependent ERK1/2 activation in arrestin-2/3 knockout MEFs. Interestingly, alanine substitution of the homologous arrestin-3 residue (lysine 308) did not significantly affect c-Raf1 binding or its ability to promote ERK1/2 activation. Together, these findings suggest that the two non-visual arrestins perform the same function via distinct molecular mechanisms. To further elucidate arrestin-MAPK interactions, we performed in vitro binding assays using pure proteins, and demonstrated that ERK2 directly binds free arrestin-2 and arrestin-3, as well as receptor-associated arrestin-1, arrestin-2, and arrestin-3. We have also shown that the arrestin-2 and arrestin-3 association with beta2-adrenergic receptors (β2ARs) significantly enhances ERK2 binding, yet has virtually no effect upon arrestins interactions with the upstream kinases c-Raf1 and MEK1.
Arrestins exist in three conformational states: free, receptor-bound, and microtubule (MT)-bound. Using conformationally biased arrestin mutants, we found that ERK2 prefers two conformations: MT-bound, mimicked by “constitutively inactive” arrestin-Δ7, and receptor-bound, mimicked by “pre-activated” arrestin-3A. Both mutants were able to rescue arrestin-mediated ERK1/2 activation in arrestin-2/3 double knockout fibroblasts. Lastly, we found that the arrestin-2 interaction with c-Raf1 is enhanced by receptor binding, whereas the interaction between arrestin-3 and c-Raf1 is not, thus suggesting that the two non-visual arrestins execute similar functions via diverse mechanisms.