Ectopic expression of the erythropoietin receptor (EpoR) in the interleukin-3 (IL-3)-dependent cell line Ba/F3 results in growth and partial erythroid differentiation in Epo. In contrast, introduction and activation of the interleukin-5 receptor (IL-5R) or of the granulocyte-macrophage colony-stimulating factor receptor (GM-CSFR) results in proliferation only. As this effect is specific to the EpoR, the role of its extracellular or cytoplasmic domain in differentiation was tested after construction of two chimeric receptors. One receptor contained the extracellular domain of EpoR fused to the endodomain of IL-3R beta-chain (E/beta), while the other contained the EpoR cytoplasmic region fused to the extracellular domain of GM-CSFR alpha-chain (GMER). Surprisingly, both receptors induced differentiation ruling out a strict specificity of the extracellular or cytoplasmic region of EpoR in this process. Instead the ability to signal differentiation correlated with structural features shared by the EpoR, GMER, and E/beta receptors. Dimerization of all three receptors results in the pairing of two signal transducing chains in the cytoplasm, in contrast to the mitogenic receptors IL-3R, IL-5R, GM-CSFR, which assemble as alphabeta heterodimers. Two new chimeric receptors that fulfilled the structural requirement exemplified by EpoR, but lacked any part of EpoR, were designed to consolidate this model. They consisted of the ectodomains of the GMR-alpha and IL-5R alpha, respectively, fused to the endodomain of IL-3R beta-chain. Both receptors were as effective as EpoR in signaling differentiation in response to their cognate ligand. Another property of receptors fulfilling these structural requirements is that they cause a marked delay in signal transducers and activators of transcription 5 (STAT5) activation on ligand stimulation. Taken together our studies show that structural assembly of receptors dictates their potential to signal erythroid differentiation in Ba/F3 cells, that differentiation can take place in the absence of Epo and that a delay in STAT5 activation is highly predictive of this process.