Proteins that contain expanded glutamine repeats are toxic to cells and lead to neurodegenerative diseases in humans — but the repeats themselves might not be entirely to blame. A study in flies and mice shows that the neuronal toxicity seen in one type of ataxia is caused by an interaction between sequences outside the glutamine tract and a protein that is involved in neuronal survival.

Spinocerebellar ataxia 1 (SCA1) is caused by the expansion of a polyglutamine (polyQ) tract in the human ataxin 1 gene (ATXN1). The associated neuronal degeneration involves a gain-of-function mechanism. However, it does not simply involve the polyQ repeats, as a version of ATXN1 that does not have a pathological number of repeats causes neuron loss when it is overexpressed in mice and flies. The fact that polyQ-expanded ATXN1 and its fly homologue ATX1 (which lacks a polyQ stretch) have similar phenotypes when overexpressed in Drosophila melanogaster meant that the authors could use the fly system as a starting point for identifying the sequences that are required for SCA1 pathogenesis and the proteins that they bind to.

Biochemical and two-hybrid data revealed that fly and human ataxin 1 bind to Senseless (SENS), a transcription factor that is involved in sensory organ development in the fly, and that the interaction involves a conserved 110 amino-acid domain (AXH) of ataxin 1. Genetic interaction studies supported the physical binding experiments and also showed that fly and human ataxin 1 cause the cell-autonomous degradation of the SENS protein. So, although the polyQ domain is not directly toxic to cells, this domain might stabilize ataxin 1 and allow it to degrade SENS and disrupt its function in sensory organ formation. A similar process occurs in mice, where human ATXN1 interacts with the mouse homologue of SENS, GFIL, through the AXH domain. In fact, in mice the destabilization of GFIL caused by ATXN1 overexpression leads to the loss of the specific neuronal types (Purkinje cells) that degenerate in patients with SCA1.

This work brings us closer to understanding how ATXN1 causes neurodegeneration. What is true for SCA1 might be true for the other eight or so polyQ-repeat-induced diseases, so studying the wild-type function of these proteins and their cellular interactions could be a fruitful strategy for identifying their pathological basis.