Phospholamban is an oligomeric transmembrane (TM) protein thoroughly studied as a contractility regulator in cardiomyopathies and as a model system for spectroscopy and protein folding. While phospholamban forms a narrow apolar channel, it debated given conflicting biochemical studies and NMR-based models whether this protein simply regulates the SERCA2 Ca2+ pump or also conducts ions. Additionally, the “Leu-Ile zipper” motif at its core poses a puzzling paradigm in TM protein folding: how does hydrophobic side chain packing drive association of alpha-helices in the already hydrophobic lipid bilayer. Meanwhile, phospholamban's X-ray crystal structure has not been reported, but would provide invaluable answers. We re-designed this protein, guided by simulation, and solved the variant's structure. First, we conducted 300 ns all-atom molecular dynamics simulations in lipid bilayers to assess the structural rigidity and pore size of the phospholamban TM domain from a reported NMR-based model. We observed the channel to widen, but remain dehydrated. The N-terminal Leu-Ile zipper motif remained very rigid (<1.5 Å RMSD), while its C-terminal polar region was dynamic and splayed open forming a water-filled pore. Our engineered variant PL5 (50% similarity) has an extended Leu-Ile zipper to replace the C-terminal polar region. In simulation, the now apolar C-terminus was rigidified and maintained the narrow dehydrated pore as previously observed. We showed PL5 is an alpha-helical coiled-coil pentamer in solution by analytical ultracentrifugation and CD spectroscopy in micelles. We then solved the X-ray crystal of PL5 in C8E4 micelles (3.17 Å), which is in striking agreement to the design model: 1.16 Å RMSD. Our data shows PL5 is a reliable model for phospholamban and the Leu-Ile zipper motif, and suggests phospholamban forms a dehydrated channel intermediate in size between the two reported NMR models unlikely to conduct ions.