DNA is a versatile tool for driving recognition-based colloidal assembly. Past work has focused on 1) employing perfectly matched sequences to form duplexes between particle surfaces and drive colloidal assembly and 2) thermal melting or dissociation of duplexes to break assemblies apart. Here, we take an alternative approach to both the assembly and disassembly process. We employ target sequences 11, 13, or 15 bases long with a single point mutation to assemble colloidal satellites. Each colloidal satellite is comprised of a large, nonfluorescent core particle surrounded by a single layer of fluorescent nanoparticles. In order to break apart or disassemble the colloidal satellites, perfectly matched 15 base-long target sequences are added to the suspension. Flow cytometry results indicate that these secondary targets can competitively displace the two shorter mismatched targets to, in principle, cause complete disassembly of colloidal satellites linked together with either mismatched sequence. Fluorescence and confocal microscopy studies, however, indicated that the extent of satellite disassembly was greatest for the 13 base-long mismatched linkages and only modest for the 11 base-long mismatched linkages. These studies thus indicate that the inclusion of mutations can cause unexpected trends in the hybridization activity between DNA-functionalized colloidal particles.