Many polypeptides overexpressed in bacteria are produced misfolded and accumulate as solid structures called inclusion bodies. Inclusion-body-prone proteins have often been reported to escape precipitation when fused to maltose-binding protein (MEP). Here, we have examined the case of HPV 16 oncoprotein E6. The unfused sequence of E6 is overexpressed as inclusion bodies in bacteria. By contrast, fusions of E6 to the C-terminus of MBP are produced soluble. We have analyzed preparations of soluble MBP-E6 fusions by using three independent approaches: dynamic light scattering, lateral turbidimetry, and sandwich ELISA. All three methods showed that MBP-E6 preparations contain highly aggregated material. The behavior of these soluble aggregates under denaturating conditions suggests that they are formed by agglomeration of misfolded E6 moieties. However, precipitation is prevented by the presence of the folded and highly soluble MBP moieties, which maintain the aggregates in solution. Therefore, the fact that a protein or protein domain is produced soluble when fused to the C-terminus of a carrier protein does not guarantee that the protein of interest is properly folded and active. We suggest that aggregation of fusion proteins should be systematically assayed, especially when these fusions are to be used for binding measurements or activity tests.