The fusion of monocyte/macrophage lineage cells into fully active, multinucleated, bone resorbing osteoclasts is a complex cell biological phenomenon that utilizes specialized proteins. OC-STAMP, a multi-pass transmembrane protein, has been shown to be required for pre-osteoclast fusion and for optimal bone resorption activity. A previously reported knockout mouse model had only mononuclear osteoclasts with markedly reduced resorption activity in vitro, but with paradoxically normal skeletal micro-CT parameters. To further explore this and related questions, we used mouse ES cells carrying a gene trap allele to generate a second OC-STAMP null mouse strain. Bone histology showed overall normal bone form with large numbers of TRAP-positive, mononuclear osteoclasts. Micro-CT parameters were not significantly different between knockout and wild type mice at 2 or 6 weeks old. At 6 weeks, metaphyseal TRAP-positive areas were lower and mean size of the areas were smaller in knockout femora, but bone turnover markers in serum were normal. Bone marrow mononuclear cells became TRAP-positive when cultured with CSF-1 and RANKL, but they did not fuse. Expression levels of other osteoclast markers, such as cathepsin K, carbonic anhydrase II, and NFATc1, were not significantly different compared to wild type. Actin rings were present, but small, and pit assays showed a 3.5-fold decrease in area resorbed. Restoring OC-STAMP in knockout cells by lentiviral transduction rescued fusion and resorption. N- and C-termini of OC-STAMP were intracellular, and a predicted glycosylation site was shown to be utilized and to lie on an extracellular loop. The site is conserved in all terrestrial vertebrates and appears to be required for protein stability, but not for fusion. Based on this and other results, we present a topological model of OC-STAMP as a 6-transmembrane domain protein. We also contrast the osteoclast-specific roles of OC- and DC-STAMP with more generalized cell fusion mechanisms.