| Transmembrane proteins make up more than a third of the kinds of cellular proteins and participate in a wide range of critical cellular processes such as cell signaling, ion transport, and oxidative phosphorylation. Helical transmembrane domains engage in highly specific interactions within the membrane and play an important role in protein oligomerization, folding, and function. However, due to their hydrophobicity, it is difficult to study transmembrane interactions in mammalian cells. The DiMaio lab has developed a genetic approach to study the structure and function of transmembrane proteins by constructing artificial transmembrane proteins that modulate cellular transmembrane proteins. Using this approach, we have isolated novel transmembrane proteins that activate the human erythropoietin receptor (hEPOR), a cytokine receptor required for red blood cell formation.;To isolate the novel hEPOR activator, we used the 44-amino acid bovine papillomavirus E5 oncoprotein as a scaffold to construct a retrovirus library expressing several hundred thousand unique hydrophobic sequences. E5 is a transmembrane protein that activates the cellular protein, the platelet-derived growth factor beta receptor (PDGFbetaR), and our goal was to re-program the viral oncoprotein to activate a different target. We screened this library of E5-like randomized sequences to select for small transmembrane proteins that conferred ligand-independent proliferation of mammalian cells expressing the hEPOR. These proteins were structurally unrelated to EPO, yet specifically activated the hEPOR and not the related murine EPOR. The most active protein, TC2-3, required an intact hEPOR transmembrane domain, induced tyrosine phosphorylation of the hEPOR, and conferred cell-autonomous growth factor independence. Most strikingly, expression of TC2-3 in human hematopoietic progenitor cells (hHPCs), which require EPO for differentiation, induced erythroid differentiation in the absence of EPO, although it was much less active than EPO. These results demonstrate that by randomizing the transmembrane domain of E5, we can re-program it so that it no longer activates its natural target and, instead, modulates an entirely different protein.;Here, we used a directed evolution approach to construct more active TC2-3 variants that would aid in mechanistic studies and allow us to explore the structural basis for the ability of transmembrane proteins to activate the hEPOR. We constructed a library that encoded several thousand TC2-3 variants, and selected a variant, EBC5-16, that displayed dramatically increased ability to confer growth factor independence in mouse cells engineered to express the hEPOR and displayed substantially increased activity in hHPCs. The only difference between EBC5-16 and TC2-3 was an isoleucine to serine substitution at position 25, which resulted in increased homodimerization. Genetic studies identified the residues that constitute the homodimer interface of EBC6-16, and biochemical analysis revealed that EBC5-16 and hEPOR formed a physical complex. We isolated a compensatory mutant suggesting that EBC5-16 and TC2-3 directly interact with the transmembrane domain of the hEPOR.;The compensatory activator, EBC2-7, required different residues in the hEPOR transmembrane domain for activity, but more significantly, it did not induce activation of the main EPOR signaling proteins, JAK2 or STAT5, indicating that it confers activity in a different manner than EBC5-16, TC2-3, and EPO. EBC2-7 co-immunoprecipitated with the point mutant receptor while EBC5-16 displayed minimal complex formation, providing additional evidence that EBC5-16 and TC2-3 directly interact with the hEPOR.;As an alternative to performing a genetic screen using several rounds of enrichment, we submitted gnomic DNA for deep sequencing after one round of selection. We identified a new hEPOR activator, BYE3, that was completely different in sequence from EBC5-16. Interestingly, BYE3 displayed decreased levels of phosphorylated JAK2 and the P13 kinase subunit, p85, and failed to induce erythroid differentiation in hHPCs. These results suggest that BYE3 activates the hEPOR in a manner distinct from EBC5-16, TC2-3, and EPO.;Through elucidating the mechanism of how these artificial proteins function, we have revealed key mechanistic insights into not only protein-protein interactions, but also hEPOR biology. These proteins may even facilitate the development of novel erythropoiesis-stimulating agents. The use of genetic selection to isolate and optimize artificial proteins is a powerful approach that can be used to modulate a variety of transmembrane proteins and also provides a potent tool to probe a diverse array of cellular activities. |
