Transmembrane Signaling by Receptor Oligomerization

The coordmation of cell growth, differentiation, and other activities in a multicellular organism is precisely controlled by a plethora of growth factors or cytokines that achieve then effects upon the cell by binding to specific cell-surface receptors. The majority of these numerous receptors for growth factors and cytokines are bitopic integral-membrane proteins that contain an extracellular ligand-binding domain; a single transmembrane domain that is assumed to be an α-helix; and a cytoplasmic-effector domain (1 ,2 ). The cytoplasmic- effector domain may have enzymatic activity, as is the case for the growth-factor receptor tyrosine kinases (1 ); or it may require interaction with other cytoplasmic-signaling molecules—notably the Janus (JAK) kinases in the case of the cytokine-receptor superfamily (2 ,3 ). Over the years, several mechanisms have been suggested for how such bitopic-membrane proteins can transmit signals across the cell membrane upon binding of their cognate ligand (4 ). Intramolecular mechanisms that have been proposed involve ligand-induced conformational changes that are propagated through the single transmembrane α-helix or alter the association of the receptor with the membrane (a “push-pull” model). Objections to these models are based upon the stability of fully hydrogen-bonded transmembrane α-helices and the ease of deformability of a lipid bilayer. Any alteration in the membrane-spanning helix is likely to be “damped” by the readily deformable membrane that it spans (see ref. 4 for a discussion). In the late 1970s studies employing fluorescence-photobleaching recovery demonstrated that several growth factors, most notably epidermal-growth factor (EGF), induce ohgomerization of their specific receptors (5 ), and that this is necessary for a biological response (6 ). Yarden and Schlessinger subsequently showed that the purified EGF receptor tyrosme kinase undergoes dimerization upon binding to EGF (7 ), and that the dimeric form of the receptor displays elevated tyrosme kinase activity (8 ). EGF-induced EGF-receptor dimerization was also demonstrated in intact cells, using chemical crosslinking approaches (9 ,10 ). As a result of these observations, a model for signal transduction by allosteric receptor oligomerization was proposed (11 ). This model has since been confirmed for many receptor tyrosine kinases in addition to the EGF receptor, as well as for many of the cytokine receptors. The general ohgomerization model holds that inactive receptor monomers are in equilibrium with active receptor dimers such that, in the absence of ligand, the eqmhbrium greatly favors the monomeric form. Upon ligand binding, the equilibrium is shifted in favor of the activated dimer (which may be a homo- or heterodliner), with resultant biological effects. In the past 10 yr, our understanding of this process has developed greatly. Where tyrosine kinase activity is a property of the receptor (the receptor tyrosme kinases) or is associated with the receptor (as with JAK kinases bound to cytokine receptors), it appears that ligand-induced receptor oligomerization brings kinase molecules into close proximity such that they can phosphorylate one another. This trans-phosphorylation, together with additional possible conformational alterations upon ohgomerization, leads to stimulation of the kinase activity—coupling receptor oligomerization to receptor activation. In this chapter, we will concentrate primarily on the mechanistic aspects of ligand-induced receptor oligomerization, selecting examples for which the process has been most thoroughly studied. A common theme emerges from these studies, in which multivalent ligand binding provides the driving force to shift the monomer/oligomer equilibrium in favor of the oligomer. There are several variations on this common theme, which appear to be exploited to enhance signal diversity for a given combination of ligands and receptors. Details of the protein-protein interactions that are involved in coupling receptor activatron to the downstream-signaling cascades are discussed in the previous two chapters by Kuriyan and Mayer, respectively.