Dentatorubral and pallidoluysian atrophy is associated with expansion of an unstable CAG repeat on chromosome 12p. We have determined the nucleotide sequences of overlapping cDNA clones and deduced the gene structure. The gene is ubiquitously expressed to form a single 4.5 kb transcript and encoded by an open reading frame of 1184 amino acids (aa), in which a polyglutamine track with variable length starts at aa 484. Although the predicted amino acid sequence does not reveal any function, it does contain several interesting motifs consisting of a simple repeated amino acid sequence, a homo-proline track, two stretches of arginine-glutamic acid dipeptides and a stretch of alternative histidine residues. These results provide clues toward understanding neurodegenerative diseases associated with triplet repeat expansion.