T. Jochum, M. E. Ritz, C. Schuster, S. F. Funderburk, K. Jehle, K. Schmitz, F. Brinkmann, M. Hirtz, D. Moss, A. C. B. Cato 

Biochim. Biophys. Acta 1822 (2012) 1070–1078 

Hormone-dependent aggregation of the androgen receptor (AR) with a polyglutamine (polyQ) stretch amplification (> 38) is considered to be the causative agent of the neurodegenerative disorder spinal and bulbar muscular atrophy (SBMA), consistent with related neurodegenerative diseases involving polyQ-extended proteins. In spite of the widespread acceptance of this common causal hypothesis, little attention has been paid to its apparent incompatibility with the observation of AR aggregation in healthy individuals with no polyQ stretch amplification. Here we used atomic force microscopy (AFM) to characterize sub-micrometer scale aggregates of the wild-type (22 glutamines) and the SBMA form (65 glutamines), as well as a polyQ deletion mutant (1 glutamine) and a variant with a normal length polyQ stretch but with a serine to alanine double mutation elsewhere in the protein. We used a baculovirus-insect cell expression system to produce full-length proteins for these structural analyses. We related the AFM findings to cytotoxicity as measured by expression of the receptors in Drosophila motoneurons or in neuronal cells in culture. We found that the pathogenic AR mutants formed oligomeric fibrils up to 300–600 nm in length. These were clearly different from annular oligomers 120–180 nm in diameter formed by the nonpathogenic receptors. We could also show that melatonin, which is known to ameliorate the pathological phenotype in the fly model, caused polyQ-extended AR to form annular oligomers. Further comparative investigation of these reproducibly distinct toxic and non-toxic oligomers could advance our understanding of the molecular basis of the polyQ pathologies.