Several neurodegenerative disorders, including Huntington's disease, are caused by expansion of the polyglutamine (polyQ) tract over 40 glutamines in the disease-related protein. Fragments of these proteins containing the expanded polyQ tract are thought to initiate aggregation and represent the toxic species. Although it is not clear how these toxic fragments are generated, in vitro data suggest that proteasomes are unable to digest polyQ tracts. To examine whether the resulting polyQ peptides could initiate aggregation in living cells, we mimicked proteasomal release of monomeric polyQ peptides. These peptides lack the commonly used starting methionine residue or any additional tag. Only expanded polyQ peptides seem to be peptidase resistant, and their accumulation initiated the aggregation process. As observed in polyQ disorders, these aggregates subsequently sequestered proteasomes, ubiquitin and polyQ proteins, and recruited Hsp70. The generated expanded polyQ peptides were toxic to neuronal cells. Our approach mimics proteasomal release of pure polyQ peptides in living cells, and represents a valuable tool to screen for proteins and compounds that affect aggregation and toxicity.