Huntington disease-like 2 (HDL2) and Huntington disease (HD) are adult-onset neurodegenerative diseases characterized by movement disorders, psychiatric disturbances and cognitive decline. Brain tissue from HD and HDL2 patients shows degeneration of the striatum and ubiquitinated inclusions immunoreactive for polyglutamine (polyQ) antibodies. Despite these similarities, the diseases result from different genetic mutations. HD is caused by a CAG repeat expansion in the huntingtin (HTT) gene, while HDL2 results from an expansion at the junctophilin 3 (JPH3) locus. Recent evidence indicates that the HDL2 expansion may give rise to a toxic polyQ protein translated from an antisense mRNA derived from the JPH3 locus. To investigate this hypothesis, we generated and characterized a Drosophila HDL2 model and compared it with a previously established HD model. We find that neuronal expression of HDL2-Q15 is not toxic, while the expression of an expanded HDL2-Q138 protein is lethal. HDL2-Q138 forms large nuclear aggregates, with only smaller puncta observed in the cytoplasm. This is in contrast to what is observed in a Drosophila model of HD, where polyQ aggregates localize exclusively to the cytoplasm. Altering localization of HLD2 with the addition of a nuclear localization or nuclear export sequence demonstrates that nuclear accumulation is required for toxicity in the Drosophila HDL2 model. Directing HDL2-Q138 to the nucleus exacerbates toxicity in multiple tissue types, while confining HDL2-Q138 to the cytoplasm restores viability to control levels. We conclude that while HD and HDL2 have similar clinical profiles, distinct pathogenic mechanisms are likely to drive toxicity in Drosophila models of these disorders.