In the second study, Harry Orr, University of Minnesota, Minneapolis, and colleagues developed new transgenic mouse models for spinocerebellar ataxia type 1 (SCA1). Their analyses revealed that wild-type ataxin-1 lacking the polyQ expansion could be made toxic by replacing a single serine with an aspartate that reshapes the protein to mimic phosphorylation at that site. Together, the new findings not only shift scientists’ thinking on how polyQ expansions cause disease, but may also help fine-tune the search for therapeutic targets for these neurodegenerative conditions.

Like Alzheimer and Parkinson diseases, polyglutamine disorders are progressive, neurodegenerative conditions marked pathologically by the clumping of proteins into inclusion bodies within the brains of affected people (Zoghbi and Orr, 2000 ). Causative mutations commonly affect proteins found within inclusions, and in polyQ diseases, the length of the glutamine repeats correlates with the extent of protein deposition and disease severity. These associations fuel the long-held view that polyQ tracts give rise to new misfolding and aggregation properties in the proteins, eventually leading to neurodegeneration. However, the data also point to a longstanding puzzle, Taylor said. Namely, why does the same mutation, unusually long glutamine stretches, lead to Huntington disease when it affects the huntingtin (Htt) gene, for instance, and to SBMA when it hits the androgen receptor?

These observations suggest that other factors must influence the toxicity of polyQ expansions. One of these is ligand binding. This idea gained traction in the SBMA field when scientists generated transgenic mice expressing full-length polyQ-containing androgen receptor (Chevalier-Larsen et al., 2004). This strain mimics the human condition, which causes motor neuron degeneration in the brainstem and spinal cord, and affects only men. However, female mice that got testosterone injections also fell prey, and males that were castrated at birth did not. These studies showed that the “androgen receptor (AR) had to bind its ligand (testosterone) to initiate the toxicity,” Taylor told ARF. “That was a revelation to me. I began to think, wait a second, maybe the polyQ toxicity is more tied to the native function of the protein than we’re giving it credit for.”

To pursue that idea, lead author Natalia Nedelsky and colleagues launched a project in Drosophila, where they expressed AR transgenes in the eye and gauged their toxicity by analyzing cell position and bristle formation in developing eye discs. By creating transgenic flies that express polyQ-expanded androgen receptors with mutations that disrupt various functional domains, the team showed that the receptor “has to do all the things it normally does, or it doesn’t cause toxicity,” Taylor said. “It has to bind ligand, move into the nucleus, bind DNA, and be able to interact with coregulator proteins through its C-terminal activation domain.” To demonstrate that last property, the researchers mutagenized this domain and found that removing the receptor’s ability to bind coregulators wiped out its pathogenic effects. “It still has 121 glutamines, which is really huge, but it’s no longer toxic,” Taylor said.

While Taylor and colleagues provide evidence to suggest that polyQ expansions are not sufficient to cause disease, Orr’s team reports in the second paper that the extra glutamines may not even be necessary. In that study, researchers led by first author Lisa Duvick generated new transgenic mouse models for SCA1, a neurodegenerative condition caused by mutations in ataxin-1 that affect motor coordination and balance. The researchers report they can get toxicity from a wild-type form of ataxin-1 with no polyQ expansion, merely by changing one serine residue (S776) to an aspartate (D776).

The work extended earlier research by Orr and co-author Huda Zoghbi, at Baylor College of Medicine in Houston, Texas, showing that ataxin-1 binds RNA-binding motif protein 17 (RBM17) and that the interactions depend on S776 and polyglutamine length. (Lim et al., 2008 and ARF news story). By replacing S776 with an aspartate that shifted the protein toward the shape it would adopt after getting phosphorylated at that site, the researchers found they could transform this otherwise wild-type ataxin-1 into one that interacted with RBM17 just as well as polyQ-expanded ataxin-1. In the current study, Duvick and colleagues tested whether this D776 mutation could cause neurodegeneration in mice.

They did this by introducing a D776 point mutation into the targeting construct Orr and Zoghbi had used previously to create transgenic mice expressing wild-type ataxin-1, with or without an expanded polyQ stretch, in cerebellar Purkinje cells, the major cell type affected in SCA1 (Burright et al., 1995 ). (A bit of alphabet soup here: bear in mind that non-expanded ataxin-1 transgenes have 30 glutamines and are denoted ATXN1[30Q]-S776 and ATXN1[30Q]-D776. The expanded forms have 82 glutamines and are denoted ATXN1[82Q]-S776 and ATXN1[82Q]-D776).

Comparing 30Q strains expressing wild-type or mutated ataxin-1 at comparable levels, the D776 mice had more atrophied dendrites, and in several tests measuring motor coordination and gait, performed just as poorly as transgenic animals expressing wild-type ataxin-1 with the 82Q expansion. Furthermore, when the two 82Q strains were analyzed side by side, the D776 mice showed a comparable extent of dendritic atrophy and neurological impairment as did S776 mice with higher transgene expression. “Thus, it took less ATXN1[82Q]-D776 than ATXN1[82Q]-S776 to induce a similar level of disease, indicating that D776 enhanced the pathogenicity of ATXN1[82Q],” the authors wrote.

Taking into account both reports, “Our thinking is moving away from the idea that the polyQ expansion causes some abnormal, new function of the protein. Rather, it corrupts the protein’s normal function,” Orr said. “Biologically, this is a major paradigm change.”

Orr’s mouse data do not suggest that expanded polyQ tracts are completely dispensable. Though the 30Q-D776 transgene seemed to drive initial pathology and motor symptoms as strongly as the 82Q-expanded form lacking the D776 mutation, 30Q-D776 mice had much less Purkinje cell atrophy than 82Q-S776 animals. Abnormalities related to native function may start disease, Taylor said, but the polyQ toxicity seem necessary to “drive it to the finish line.”

These papers fuel a trend that calls into question the prevailing view that polyQ expansions cause disease, Orr said. “They may play a role in the end-stage aspects of disease. But from the standpoint of curing the disease, if you simply block formation of aggregates, it's unlikely the patient is going to see any effect, as long as you keep the mutant protein.” Other research has put forth the idea that HD and PD inclusions may in fact be protective (Bodner et al., 2006 and ARF news story) and that insoluble protein deposits accumulate during the lifespan of normal C. elegans worms (David et al., 2010 and ARF news story).

If polyQ expansions in proteins do not always promote disease, what do they do? Structural studies on Htt, the protein that underlies Huntington disease, suggest that the three-dimensional conformation of the protein determines the extent of toxicity (Nekooki-Machida et al., 2009 and ARF news story) and that the polyQ stretch may dictate which conformation Htt adopts (Kim et al., 2009 and ARF news story).

Shifting protein conformation, in turn, has biological consequences, as explained in an accompanying Neuron commentary by Ian Kratter and Steven Finkbeiner of the Gladstone Institute of Neurological Disease in San Francisco. “The host protein might adopt one of several conformations that depend on specific post-translational modifications, e.g. phosphorylation, with different conformations associated with particular functions,” they write. “In the disease setting, polyQ expansion could stabilize certain conformations of the host protein at the expense of others, disrupting the equilibrium of protein conformations and enhancing the functions associated with certain conformations.”

In SBMA, the androgen receptor has been shown to exist in distinct multimeric complexes (see review Heemers and Tindall, 2007), and Taylor suspects that polyQ length influences the distribution of the receptor among them. “PolyQ proteins are multi-functional, and the polyQ expansion influences the balance of the different activities the protein is involved in,” he said.

Furthermore, Orr said, understanding the normal function of the proteins underlying SBMA and SCA seems to be “giving us some good targets for therapeutic development.” For SBMA, such discoveries have fueled efforts to disrupt antigen binding to receptor. “In the case of SCA1, a key target now is the S776 phosphorylation step,” Orr said. He suggested there may be a similar benefit in AD if researchers had a better handle on “the normal function of the pathway resulting in cleavage of amyloid precursor protein (APP) into different fragments, and how mutations in APP may skew these normal functional pathways.”—Esther Landhuis.

Reference:
Nedelsky NB, Pennuto M, Smith RB, Palazzolo I, Moore J, Nie Z, Neale G, Taylor JP. Native functions of the androgen receptor are essential to pathogenesis in a Drosophila model of spinobulbar muscular atrophy. Neuron. 23 September 2010. Abstract

Duvick L, Barnes J, Ebner B, Agrawal S, Andresen M, Lim J, Giesler GJ, Zoghbi HY, Orr HT. SCA1-like disease in mice expressing wild-type ataxin-1 with a serine to aspartic acid replacement at residue 776. Neuron. 23 September 2010. Abstract

Kratter and Finkbeiner. PolyQ Disease: Too Many Qs, Too Much Function? Neuron. 23 September 2010. Abstract