Mutant proteins that contain stretches called polyQ repeats can misfold or form aggregates linked to neurodegeneration. It emerges that some polyQ-containing proteins regulate a process that degrades misfolded proteins. See Letter p.108
Nine inherited neurodegenerative disorders1 are associated with genes that contain repetitive expansions of the DNA nucleotide sequence CAG in the normal gene sequence. CAG encodes the amino acid glutamine (termed Q in the single-letter amino-acid code), and this type of sequence expansion gives rise to proteins that have longer than usual, continuous stretches of glutamine, known as polyQ regions1. However, the normal role of these polyQ repeats is unknown. It is thought that when these stretches are longer than normal in a given protein, they can interact to form aggregates that can bind and sequester many essential cellular proteins. Such protein aggregation is associated with neuronal-cell death1.
On page 108, Ashkenazi et al.2 report a previously unknown connection between polyQ-region expansions and abnormalities in an intracellular degradation and recycling process known as autophagy. The discovery of a link between the presence of polyQ regions and the regulation of autophagy might indicate that polyQ-containing proteins have a role in normal cellular homeostasis that is distinct from their action in disease.
Autophagy removes misfolded proteins and damaged organelles by encapsulating them in a lipid membrane compartment called an autophagosome. This then fuses with an acidic, membrane-bound organelle known as a lysosome to form an autophagolysosome, in which the damaged or malfunctioning substrates are degraded3. The development of treatments for diseases associated with polyQ expansions is currently aimed at reducing the level of the disease-causing proteins, and promotion of autophagy has emerged as a promising therapeutic goal3.
Expansions of polyQ domains in the proteins huntingtin and ataxin 3 can cause Huntington's disease and spinocerebellar ataxia type 3, respectively1. In the former disease, the loss of neurons in a region of the brain called the striatum leads to changes in mood and personality, as well as the loss of motor coordination and the development of involuntary movements3. In spinocerebellar ataxia type 3, neurodegeneration occurs in the striatum and in a brain region known as the cerebellum, resulting in loss of motor coordination. Both diseases are of a progressive nature, and the age of disease onset is inversely related to polyQ length. No therapies are currently available to slow the progression of diseases associated with the presence of expanded polyQ regions.
Ashkenazi et al. show that the addition of a protein tag called ubiquitin to a key regulator of autophagy initiation, the protein beclin 1, marks beclin 1 for destruction in the proteasome degradation complex. However, removal of the ubiquitin tag by the enzyme ataxin 3 allows beclin 1 to escape proteasomal destruction and trigger autophagy. Importantly, the authors found that the interaction between ataxin 3 and beclin 1 can be prevented if other proteins containing expanded polyQ domains are more successful at interacting with beclin 1 than is ataxin 3 (Fig. 1). They show that a mutated version of huntingtin that contains an expanded polyQ region can compete with ataxin 3 for binding to beclin 1. Huntingtin binding to beclin 1 would prevent ataxin 3 from binding, and therefore promote beclin 1 degradation by the proteasome. This finding suggests that dysregulated autophagy in Huntington's disease might be due, at least in part, to an increase in beclin 1 degradation.
Ashkenazi and colleagues report that cells from people who have certain polyQ-associated diseases had low levels of beclin 1. The authors found that although mutated ataxin 3 containing an expanded polyQ region binds beclin 1 more strongly than does normal ataxin 3, the mutated protein shows poor activity in removing ubiquitin, leading to increased beclin 1 degradation. Overall, these findings provide a framework for understanding polyQ-associated diseases, by considering these conditions as being influenced by the interplay between beclin 1 and proteins that have expanded stretches of polyQ domains.
Nevertheless, much remains to be discovered about how other proteins outcompete normal ataxin 3 for binding to beclin 1. Can this occur over a wide range of lengths of polyQ expansion, or is a specific minimum length needed for beclin 1 binding? How such a phenomenon plays out in each polyQ-associated disease needs to be resolved, particularly with regard to people in whom the lengths of polyQ repeats are intermediate between those associated with normal function and those associated with disease. Would intermediate-sized polyQ proteins put individuals at an increased risk of decreased autophagy, especially in the presence of other genetic or environmental factors that might enhance such an effect?
Ashkenazi and colleagues' work adds to the emerging complexity of huntingtin's multifaceted role in the regulation of autophagy, and provides insight into the growing list of autophagy-associated domains identified in huntingtin4,5,6,7. For example, the protein has been shown4 to participate in stress-activated autophagy by competing for binding to an autophagic regulator protein that lies upstream of beclin 1 activity.
The authors show that deletion of the polyQ region and the polyQ-binding domain do not completely block the interaction between beclin 1 and ataxin 3 or huntingtin, suggesting that other domains might be involved. This could partly explain the authors' unexpected finding that starvation-induced autophagy (fasting is a common inducer of autophagy) was inhibited in a mouse model of Huntington's disease in which cells expressed only a short fragment of huntingtin. By contrast, two different fasting regimes induced autophagy in the brain of model mice for Huntington's disease in which full-length huntingtin was expressed8. This suggests that full-length huntingtin might be required for correct activation of autophagy. Clarifying how wild-type huntingtin and other polyQ proteins interact with beclin 1 will be instrumental in fully understanding the link between these proteins and autophagy.
At a cellular level, the interrelationship between the various huntingtin domains associated with autophagy must be elucidated, to establish whether these domains regulate autophagy independently or in a coordinated fashion. It is possible that other naturally occurring variants in wild-type huntingtin, or in other polyQ-containing proteins, influence autophagy in disease settings.
The authors' study opens up several avenues for research. It indicates that enhanced autophagy might be beneficial in specific disease states. If so, it will be useful to establish whether there is a specific therapeutic window for when and for how long autophagy should be induced. Identification of the enzyme that adds the ubiquitin tag to beclin 1, and understanding how this protein is regulated, might provide therapeutic targets relevant to other polyQ-expansion-associated diseases.
Finally, it will be essential to determine how modulation of autophagy could be of use in treating people with polyQ-linked diseases, and in which situations its enhancement might be beneficial. For example, a clinical trial is under way9 in which people with Huntington's disease are receiving a drug that reduces the expression of both the wild-type and mutant copies of the gene encoding huntingtin. How will these people be affected by losing the benefit of wild-type huntingtin in autophagy? Alternatively, could induction of autophagy in response to fasting complement a treatment that is specific for mutant huntingtin? It is clear that understanding and potentially harnessing the protective effects of autophagy in the treatment of polyQ-associated diseases warrants further investigation. Footnote 1
Menzies, F. M. et al. Neuron 93, 1015–1034 (2017).
Ashkenazi, A. et al. Nature 545, 108–111 (2017).
Martin, D. D. O., Ladha, S., Ehrnhoefer, D. E. & Hayden, M. R. Trends Neurosci. 38, 26–35 (2015).
Rui, Y.-N. et al. Nature Cell Biol. 17, 262–275 (2015).
Ochaba, J. et al. Proc. Natl Acad. Sci. USA 111, 16889–16894 (2014).
Wong, Y. C. & Holzbaur, E. L. F. J. Neurosci. 34, 1293–1305 (2014).
Martin, D. D. O. et al. Hum. Mol. Genet. 23, 3166–3179 (2014).
Ehrnhoefer, D. E. et al. Preprint at bioRxiv https://doi.org/10.1101/116178 (2017).
About this article
Science Signaling (2017)