Epidermal growth factor receptor (EGFR) ligands undergo a complex series of processing events during their maturation to active signaling proteins. Like its mammalian homologs, the predominant Drosophila EGFR ligand Spitz is produced as a transmembrane pro-protein. In the secretory pathway, Spitz is cleaved within its transmembrane domain to release the extracellular signaling domain. This domain is modified with an N-terminal palmitate group that tethers it to the plasma membrane. We found that the pro-protein can reach the cell surface in the absence of proteolysis, but that it fails to activate the EGFR. To address why the transmembrane pro-protein is inactive, whereas membrane association through the palmitate group promotes activity, we generated a panel of chimeric constructs containing the Spitz extracellular region fused to exogenous transmembrane proteins. Although the orientation of the EGF domain and its distance from the plasma membrane varies in these chimeras, they are all active in vivo. Thus, tethering Spitz to the membrane via a transmembrane domain at either terminus does not prevent activity. Conversely, removing the N-terminal palmitate group from the C-terminally tethered pro-protein does not render it active. Furthermore, we show that the Spitz transmembrane pro-protein can activate the EGFR in a tissue culture assay, indicating that its failure to signal in vivo is not due to structural features. In polarized imaginal disc cells, unprocessed Spitz pro-protein localizes to apical puncta, whereas the active chimeric Spitz constructs are basolaterally localized. Taken together, our data support the model that localized trafficking of the pro-protein restricts its ability to activate the receptor in polarized tissues.