news and views b
Late in session
Early in session
Marina Corral Spence/Nature Publishing Group
a
npg
© 2014 Nature America, Inc. All rights reserved.
Figure 1 Studying learning using an optical BCI. (a) Schematic of the closed-loop BCI system. Neural activity, measured by optical imaging of an intracellular calcium indicator, is converted in real time to auditory feedback (tone pitch). A mouse is trained to modulate neural activity to attain a specified pitch level for reward. (b) In a basic version of this task, the pitch is directly related to the activity of one neuron (, dark blue) minus the activity of another neuron (, dark red), together referred to as the direct neurons. Box indicates field of view of two-photon microscope; each circle represents a neuron. Initially, the nearby indirect neurons (light blue and light red) change their activity together with the direct neurons in a distance-dependent manner (left). Over the course of a few tens of minutes, this correlation decreases, suggesting that the increase in performance seen during a single session is tied to the dissociation of the activity of direct and indirect neurons (right).
for basic scientific investigation of learning. Optical imaging has several advantages over electrode recordings, including providing a finer spatial resolution and allowing investigation of most neurons in a field of view. It can also provide a more unbiased view of the neural population, including neurons with low firing rates that may be missed using electrode recordings. Furthermore, optical recordings may facilitate the assessment of cell type and anatomical connectivity among the recorded neurons13,14. Conversely, it is easier to monitor deep brain structures with electrical recordings, as light scattering can limit recording resolution when attempting to penetrate far into neural tissue. Electrical recordings also provide finer temporal resolution than optical imaging, which is necessary if one seeks to study precise spike timing3,15. By working with mice, one can take advantage of the impressive array of genetic tools to selectively record from and manipulate the activity of subpopulations of neurons. This should allow for a careful dissection of the specific circuit mechanisms that underlie the
learning of new behaviors. This new approach complements ongoing BCI work in nonhuman primates, for which most of these genetic tools are not now available. In nonhuman primates, however, it is easier to scale up the richness of the task from a one-dimensional auditory tone to a two-dimensional computer cursor or a multi-dimensional robotic arm. For basic science, this is important because richer BCI tasks are more likely than simpler BCI tasks to engage the complex, cognitive processes that underlie native arm control and other behaviors. Furthermore, for rehabilitation purposes, richer BCI tasks in nonhuman primates provide a more direct translation route to systems used with patients5,6. Whereas football fans can learn whether to flash up cards by determining where they are sitting in the stadium, the neural circuits and mechanisms that underlie the learning of new behaviors are still incompletely understood. The tool provided by Clancy et al.1 enables a new view of the neural circuit, complementing existing tools for investigating the neural basis of learning. The field is now at a point where neuroscientists
have a choice among powerful techniques for studying learning phenomena at multiple levels, from the subcellular through the neuronal to the circuit and systems levels. Go team! COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Clancy, K.B., Koralek, A.C., Costa, R.M., Feldman, D.E & Carmena, J.M. Nat. Neurosci. 17, 807–809 (2014). 2. Scanziani, M. & Hausser, M. Nature 461, 930–939 (2009). 3. Koralek, A.C., Costa, R.M. & Carmena, J.M. Neuron 79, 865–872 (2013). 4. Ganguly, K., Dimitrov, D.F., Wallis, J.D. & Carmena, J.M. Nat. Neurosci. 14, 662–667 (2011). 5. Collinger, J.L. et al. Lancet 381, 557–564 (2013). 6. Hochberg, L.R. et al. Nature 485, 372–375 (2012). 7. Jarosiewicz, B. et al. Proc. Natl. Acad. Sci. USA 105, 19486–19491 (2008). 8. Chase, S.M., Kass, R.E. & Schwartz, A.B. J. Neurophysiol. 108, 624–644 (2012). 9. Komiyama, T. et al. Nature 464, 1182–1186 (2010). 10. Gilja, V. et al. Nat. Neurosci. 15, 1752–1757 (2012). 11. Golub, M., Yu, B.M. & Chase, S.M. Conf. Proc. IEEE Eng. Med. Biol. Soc. 2012,1327–1330 (2013). 12. Wander, J.D. & Rao, R.P. Curr. Opin. Neurobiol. 25C, 70–75 (2014). 13. Ko, H. et al. Nature 496, 96–100 (2013). 14. Bock, D.D. et al. Nature 471, 177–182 (2011). 15. Engelhard, B., Ozeri, N., Israel, Z., Bergman, H. & Vaadia, E. Neuron 77, 361–375 (2013).
Impaired import: how huntingtin harms Elizabeth A Jonas We now learn that mutant huntingtin binds to a complex that imports constituent proteins across the mitochondrial inner membrane, halting bioenergetics in synaptic mitochondria and predisposing to neuronal dysfunction and death. Mitochondria are the wellspring of life, increasingly heralded as necessary for the Elizabeth A. Jonas is in the Departments of Internal Medicine and Neurobiology, Yale University, New Haven, Connecticut, USA. e-mail: [email protected]
revention of degeneration of vulnerable popp ulations of neurons. Mitochondrial metabolism is crucial to neuronal function1, and compromised mitochondria have been found in neurodegenerating cells2, with mitochondrial dysfunction being implicated in every major neurodegenerative disorder, including
nature neuroscience volume 17 | number 6 | June 2014
Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis and Huntington’s disease3. Whether mitochondrial dysfunction is the sole cause of the disorder or secondary to other deficits has continued to be debated, but understanding the implications of this question is important to determining the 747
underpinnings of neurodegeneration. In this issue of Nature Neuroscience, Yano et al.4 describe a breakthrough in understanding the link between mitochondrial dysfunction and neuronal death in Huntington’s disease. They find that mutant huntingtin protein causes a specific blockade of the mechanism that provides mitochondria with almost all of the proteins they need to run metabolism on a daily basis. A classic hallmark of neurodegenerative disease is abnormal accumulation of distorted proteins. In Huntington’s disease, patients may dominantly inherit an aberrant form of the gene huntingtin (HTT) that contains an expanded sequence of CAG repeats resulting in an extra-long string of glutamines in the huntingtin protein5. The function of normal huntingtin is unknown, but it is clear that the abnormal protein aggregates inside degenerating striatal neurons close to mitochondria 6. Although human tissue and mouse models have demonstrated that aggregated huntingtin proteins do not directly cause the death of striatal neurons7, death may be a result of interference of mutant huntingtin with mitochondrial calcium handling8 or metabolic functions9. Indeed, in addition to the severe motor and cognitive deficits found in Huntington’s disease, patients also display weight loss leading to early death10. How could mutant huntingtin protein cause mitochondrial dysfunction? Recent work on brain samples from patients with Huntington’s disease and in mouse models has revealed reduced enzymatic activity of members of the inner mitochondrial membrane electron transport chain, the main enzyme system that produces the cell’s energy supply9. This finding is correlated with reduced levels of ATP in the mutant neurons and reduced uptake of substrates such as glucose by mitochondria and by the cells affected by the mutant protein. Mutant mitochondria also display abnormal morphology and calcium handling, decreased movement, and a tendency to fragment11. All of these defects are consistent with the notion that mitochondria are starved for the building blocks necessary for ongoing maintenance and repair. These repair processes include replacement of damaged proteins. How do mitochondria normally replace proteins? Mitochondria were, in ancient times, separate prokaryotic cells that incorporated themselves into eukaryotic cells and became symbiotes. They gave up most of their DNA to the more efficient host cell nuclear genome. Mitochondria now contain DNA for just 37 genes, mostly encoding 748
subunit members of the electron transport chain, mitochondrial ribosomal proteins and tRNA12. This means that, to carry out their vital tasks, mitochondria continually import over 1,200 proteins from the cytosol13. After a preprotein is made in the cytosol, it must translocate through two mitochondrial membranes if it is to reach the mitochondrial matrix. Each preprotein has a specific signal or targeting sequence that allows it to bind to a group of mitochondrial proteins known as translocators. There is a general translocator in the outer mitochondrial membrane (TOM complex) and at least two types of translocator in the inner mitochondrial membrane (TIM complexes) (Fig. 1). These translocators work together to take up new proteins from the cytosol and sort them so they end up at their appropriate locations in the mitochondrion14.
Yano et al.4 confirmed that mutant huntingtin protein colocalizes with mitochondria. Their surprising new finding, however, is that mutant huntingtin specifically binds to one of the inner membrane translocator complexes (Fig. 1). The investigators applied recombinant tagged N-terminal fragments of mutant huntingtin or normal huntingtin to isolated brain mitochondria, then immunoprecipitated the fragments and used mass spectroscopy to identify their protein-binding partners. Several subunits of the inner membrane translocator complex called TIM23 were identified among the targets that were bound more tightly by the mutant protein than by the wild-type protein, and other import complexes were not on the list, suggesting specificity in the relationship between mutant huntingtin and TIM23. TIM23 is responsible for importing
Cytosol Mitochondrial outer membrane
TOM complex
Intermembrane space TIM23 complex
50 23
Mitochondrial inner membrane
23 17
17 44
44 Matrix
Matrix, inner and intermembrane polypeptides
Mutant huntingtin protein PolyQ
Mutant huntingtin peptide N
N Figure 1 Mutant huntingtin inhibits mitochondrial protein import. Shown are the cytosolic compartment, mitochondrial inner and outer membranes, intermembrane space, and mitochondrial matrix. Mitochondrial preproteins translocate through the TOM complex in the outer membrane. Matrix and many inner membrane proteins are recognized by TIM23 subunits and translocate through the TIM23 complex. Mutant huntingtin full-length protein or mutant huntingtin N-terminal peptide, both containing long polyglutamine (polyQ) repeat sequences, bind the portion of the TIM23 complex consisting of Tim23, Tim17a and Tim50 (magenta), inhibiting protein import.
volume 17 | number 6 | June 2014 nature neuroscience
Marina Corral Spence/Nature Publishing Group
npg
© 2014 Nature America, Inc. All rights reserved.
news and views
npg
© 2014 Nature America, Inc. All rights reserved.
news and views proteins destined for the matrix or inner membrane, including many members of complexes that carry out intermediary metabolism and respiration. If mutant huntingtin is stuck to the TIM23 complex, it might disrupt protein import into the matrix or inner membrane, severely hampering metabolism in cells such as neurons that are highly energy dependent. Using an established assay, the investigators found that mutant huntingtin N-terminal peptides inhibited protein import into mitochondria isolated from mouse brain in a manner dependent on the concentration and length of the huntingtin polyglutamine sequence. The authors then tested protein import in vivo in a mouse model of Huntington’s disease. The R6/2 mouse expresses an N-terminal fragment from a human HTT gene and has become an important model because its disease progression mimics the motor abnormalities, metabolic disarray and shortened life span found in human patients10. Perhaps suspecting the importance of synaptic dysfunction in neurodegeneration, the authors isolated highly purified synaptosomal mitochondria (those residing near presynaptic terminals) in addition to non-synaptosomal mitochondria from the mutant mice. Synaptosomal mitochondria are a more select group. They are taken only from cells with synapses—that is, neurons— and may also function differently from mitochondria isolated from the soma because these specialized organelles buffer calcium and provide energy for synaptic activity. Because there are many demands placed on synaptic mitochondria, the authors guessed that their function might become more easily compromised by the mutant protein. Notably, the authors found that the synaptic mitochondria exhibited markedly compromised protein import even at presymptomatic and early disease stages (when the mice were 3 weeks old), whereas nonsynaptic mitochondria were protected from a decline in protein import even as late as 10–11 weeks of life. Mitochondria from liver also showed no deficit in protein import until much later in the disease, suggesting that synaptic mitochondria are particularly sensitive to these mutations. Even mitochondria isolated from embryonic neurons made from the R6/2 mouse before birth displayed the protein
import defect, suggesting that this deficit is one of the earliest manifestations of disease in mutant mouse synapses. If protein import into mitochondria is needed for normal mitochondrial respiration, then the delay in protein import should precede the onset of any respiratory chain dysfunction. The authors found that there were no deficits in respiratory function in mitochondria isolated from presymptomatic animals, even though these mitochondria showed signs of severe dysfunction in protein import, whereas respiratory function became quite compromised later in the disease. Of particular note, although the deficits in respiratory chain function have previously been attributed to high levels of reactive oxygen species (ROS) causing protein and DNA damage to vulnerable mitochondria, and this may still be true, Yano et al.4 show that sublethal doses of ROS, although further damaging protein import in mitochondria from mutant huntingtin–expressing cells, do not damage the protein import pathway in normal mitochondria, ruling out ROS as the primary cause of the protein import deficit. Is the change in protein import itself sufficient to produce neuronal dysfunction and death? To address this, the authors bypassed the mutant huntingtin protein to study the effect of depletion of the general import machinery directly. They first depleted the outer membrane translocator protein TOM40. They noted that cells died by apoptotic and non-apoptotic mechanisms soon after halting protein import. Next, the authors depleted the actual target of mutant huntingtin, the inner membrane translocator component TIM23. In live neurons, they were able to watch the loss of fluorescence of a voltage-dependent mitochondrial membrane potential indicator and the accumulation of a fluorescent cell death marker as the levels of TIM23 declined. Neuronal death occurred rapidly after loss of mitochondrial membrane potential, within 1–2 days of the onset of TIM23 depletion. The authors also found that increasing TIM23 in mutant huntingtin–expressing neurons produced an improvement in metabolic health and in neuronal survival. These striking findings—that reduced mitochondrial protein import caused by mutant huntingtin leads to metabolic dysfunction and death of neurons—raise
nature neuroscience volume 17 | number 6 | June 2014
many questions. Why do some neurons die and others survive if mutant huntingtin is expressed in all neurons? Is the amount of mutant huntingtin bound to the import machinery different in different cell populations, and, if so, how does this come about? And what about the intriguing finding that synaptic mitochondrial protein import is more affected by mutant huntingtin than that of somatic mitochondria? Is this discrepancy caused by a different level of activity of these mitochondria, increasing their vulnerability through an increase in reactive oxygen species? In turn, do these defective synaptic mitochondria cause some of the motor and cognitive symptoms of patients or animals found preceding cell death? If the protein import deficit causes neuronal death so rapidly, why do most patients develop neuronal loss and clinical disease late in life, and why is there often no neuronal loss in animal models? Even though there are still so many unanswered questions, the findings by Yano et al. 4 that mutant huntingtin causes loss of mitochondrial protein import, resulting in a decline in metabolism and respiration, certainly provides an influx of ideas to the neuro degeneration story. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Harris, J.J., Jolivet, R. & Attwell, D. Neuron 75, 762–777 (2012). 2. Petrozzi, L., Ricci, G., Giglioli, N.J., Siciliano, G. & Mancuso, M. Biosci. Rep. 27, 87–104 (2007). 3. Chaturvedi, R.K. & Beal, M.F. Ann. NY Acad. Sci. 1147, 395–412 (2008). 4. Yano, H. et al. Nat. Neurosci. 17, 822–831 (2014). 5. Kumar, A., Vaish, M. & Ratan, R.R. Drug Discov. Today doi:10.1016/j.drudis.2014.03.016 (21 March 2014). 6. Oliveira, J.M. J. Neurochem. 114, 1–12 (2010). 7. Kuemmerle, S. et al. Ann. Neurol. 46, 842–849 (1999). 8. Brustovetsky, N. et al. J. Neurosci. 23, 4858–4867 (2003). 9. Grünewald, T. & Beal, M.F. Ann. NY Acad. Sci. 893, 203–213 (1999). 10. Crook, Z.R. & Housman, D. Neuron 69, 423–435 (2011). 11. Reddy, P.H. & Shirendeb, U.P. Biochim. Biophys. Acta 1822, 101–110 (2012). 12. Anderson, S. et al. Nature 290, 457–465 (1981). 13. Calvo, S.E. & Mootha, V.K. Annu. Rev. Genomics Hum. Genet. 11, 25–44 (2010). 14. Rehling, P., Wiedemann, N., Pfanner, N. & Truscott, K.N. Crit. Rev. Biochem. Mol. Biol. 36, 291–336 (2001).
749
Late in session
Early in session
Marina Corral Spence/Nature Publishing Group
a
npg
© 2014 Nature America, Inc. All rights reserved.
Figure 1 Studying learning using an optical BCI. (a) Schematic of the closed-loop BCI system. Neural activity, measured by optical imaging of an intracellular calcium indicator, is converted in real time to auditory feedback (tone pitch). A mouse is trained to modulate neural activity to attain a specified pitch level for reward. (b) In a basic version of this task, the pitch is directly related to the activity of one neuron (, dark blue) minus the activity of another neuron (, dark red), together referred to as the direct neurons. Box indicates field of view of two-photon microscope; each circle represents a neuron. Initially, the nearby indirect neurons (light blue and light red) change their activity together with the direct neurons in a distance-dependent manner (left). Over the course of a few tens of minutes, this correlation decreases, suggesting that the increase in performance seen during a single session is tied to the dissociation of the activity of direct and indirect neurons (right).
for basic scientific investigation of learning. Optical imaging has several advantages over electrode recordings, including providing a finer spatial resolution and allowing investigation of most neurons in a field of view. It can also provide a more unbiased view of the neural population, including neurons with low firing rates that may be missed using electrode recordings. Furthermore, optical recordings may facilitate the assessment of cell type and anatomical connectivity among the recorded neurons13,14. Conversely, it is easier to monitor deep brain structures with electrical recordings, as light scattering can limit recording resolution when attempting to penetrate far into neural tissue. Electrical recordings also provide finer temporal resolution than optical imaging, which is necessary if one seeks to study precise spike timing3,15. By working with mice, one can take advantage of the impressive array of genetic tools to selectively record from and manipulate the activity of subpopulations of neurons. This should allow for a careful dissection of the specific circuit mechanisms that underlie the
learning of new behaviors. This new approach complements ongoing BCI work in nonhuman primates, for which most of these genetic tools are not now available. In nonhuman primates, however, it is easier to scale up the richness of the task from a one-dimensional auditory tone to a two-dimensional computer cursor or a multi-dimensional robotic arm. For basic science, this is important because richer BCI tasks are more likely than simpler BCI tasks to engage the complex, cognitive processes that underlie native arm control and other behaviors. Furthermore, for rehabilitation purposes, richer BCI tasks in nonhuman primates provide a more direct translation route to systems used with patients5,6. Whereas football fans can learn whether to flash up cards by determining where they are sitting in the stadium, the neural circuits and mechanisms that underlie the learning of new behaviors are still incompletely understood. The tool provided by Clancy et al.1 enables a new view of the neural circuit, complementing existing tools for investigating the neural basis of learning. The field is now at a point where neuroscientists
have a choice among powerful techniques for studying learning phenomena at multiple levels, from the subcellular through the neuronal to the circuit and systems levels. Go team! COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Clancy, K.B., Koralek, A.C., Costa, R.M., Feldman, D.E & Carmena, J.M. Nat. Neurosci. 17, 807–809 (2014). 2. Scanziani, M. & Hausser, M. Nature 461, 930–939 (2009). 3. Koralek, A.C., Costa, R.M. & Carmena, J.M. Neuron 79, 865–872 (2013). 4. Ganguly, K., Dimitrov, D.F., Wallis, J.D. & Carmena, J.M. Nat. Neurosci. 14, 662–667 (2011). 5. Collinger, J.L. et al. Lancet 381, 557–564 (2013). 6. Hochberg, L.R. et al. Nature 485, 372–375 (2012). 7. Jarosiewicz, B. et al. Proc. Natl. Acad. Sci. USA 105, 19486–19491 (2008). 8. Chase, S.M., Kass, R.E. & Schwartz, A.B. J. Neurophysiol. 108, 624–644 (2012). 9. Komiyama, T. et al. Nature 464, 1182–1186 (2010). 10. Gilja, V. et al. Nat. Neurosci. 15, 1752–1757 (2012). 11. Golub, M., Yu, B.M. & Chase, S.M. Conf. Proc. IEEE Eng. Med. Biol. Soc. 2012,1327–1330 (2013). 12. Wander, J.D. & Rao, R.P. Curr. Opin. Neurobiol. 25C, 70–75 (2014). 13. Ko, H. et al. Nature 496, 96–100 (2013). 14. Bock, D.D. et al. Nature 471, 177–182 (2011). 15. Engelhard, B., Ozeri, N., Israel, Z., Bergman, H. & Vaadia, E. Neuron 77, 361–375 (2013).
Impaired import: how huntingtin harms Elizabeth A Jonas We now learn that mutant huntingtin binds to a complex that imports constituent proteins across the mitochondrial inner membrane, halting bioenergetics in synaptic mitochondria and predisposing to neuronal dysfunction and death. Mitochondria are the wellspring of life, increasingly heralded as necessary for the Elizabeth A. Jonas is in the Departments of Internal Medicine and Neurobiology, Yale University, New Haven, Connecticut, USA. e-mail: [email protected]
revention of degeneration of vulnerable popp ulations of neurons. Mitochondrial metabolism is crucial to neuronal function1, and compromised mitochondria have been found in neurodegenerating cells2, with mitochondrial dysfunction being implicated in every major neurodegenerative disorder, including
nature neuroscience volume 17 | number 6 | June 2014
Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis and Huntington’s disease3. Whether mitochondrial dysfunction is the sole cause of the disorder or secondary to other deficits has continued to be debated, but understanding the implications of this question is important to determining the 747
underpinnings of neurodegeneration. In this issue of Nature Neuroscience, Yano et al.4 describe a breakthrough in understanding the link between mitochondrial dysfunction and neuronal death in Huntington’s disease. They find that mutant huntingtin protein causes a specific blockade of the mechanism that provides mitochondria with almost all of the proteins they need to run metabolism on a daily basis. A classic hallmark of neurodegenerative disease is abnormal accumulation of distorted proteins. In Huntington’s disease, patients may dominantly inherit an aberrant form of the gene huntingtin (HTT) that contains an expanded sequence of CAG repeats resulting in an extra-long string of glutamines in the huntingtin protein5. The function of normal huntingtin is unknown, but it is clear that the abnormal protein aggregates inside degenerating striatal neurons close to mitochondria 6. Although human tissue and mouse models have demonstrated that aggregated huntingtin proteins do not directly cause the death of striatal neurons7, death may be a result of interference of mutant huntingtin with mitochondrial calcium handling8 or metabolic functions9. Indeed, in addition to the severe motor and cognitive deficits found in Huntington’s disease, patients also display weight loss leading to early death10. How could mutant huntingtin protein cause mitochondrial dysfunction? Recent work on brain samples from patients with Huntington’s disease and in mouse models has revealed reduced enzymatic activity of members of the inner mitochondrial membrane electron transport chain, the main enzyme system that produces the cell’s energy supply9. This finding is correlated with reduced levels of ATP in the mutant neurons and reduced uptake of substrates such as glucose by mitochondria and by the cells affected by the mutant protein. Mutant mitochondria also display abnormal morphology and calcium handling, decreased movement, and a tendency to fragment11. All of these defects are consistent with the notion that mitochondria are starved for the building blocks necessary for ongoing maintenance and repair. These repair processes include replacement of damaged proteins. How do mitochondria normally replace proteins? Mitochondria were, in ancient times, separate prokaryotic cells that incorporated themselves into eukaryotic cells and became symbiotes. They gave up most of their DNA to the more efficient host cell nuclear genome. Mitochondria now contain DNA for just 37 genes, mostly encoding 748
subunit members of the electron transport chain, mitochondrial ribosomal proteins and tRNA12. This means that, to carry out their vital tasks, mitochondria continually import over 1,200 proteins from the cytosol13. After a preprotein is made in the cytosol, it must translocate through two mitochondrial membranes if it is to reach the mitochondrial matrix. Each preprotein has a specific signal or targeting sequence that allows it to bind to a group of mitochondrial proteins known as translocators. There is a general translocator in the outer mitochondrial membrane (TOM complex) and at least two types of translocator in the inner mitochondrial membrane (TIM complexes) (Fig. 1). These translocators work together to take up new proteins from the cytosol and sort them so they end up at their appropriate locations in the mitochondrion14.
Yano et al.4 confirmed that mutant huntingtin protein colocalizes with mitochondria. Their surprising new finding, however, is that mutant huntingtin specifically binds to one of the inner membrane translocator complexes (Fig. 1). The investigators applied recombinant tagged N-terminal fragments of mutant huntingtin or normal huntingtin to isolated brain mitochondria, then immunoprecipitated the fragments and used mass spectroscopy to identify their protein-binding partners. Several subunits of the inner membrane translocator complex called TIM23 were identified among the targets that were bound more tightly by the mutant protein than by the wild-type protein, and other import complexes were not on the list, suggesting specificity in the relationship between mutant huntingtin and TIM23. TIM23 is responsible for importing
Cytosol Mitochondrial outer membrane
TOM complex
Intermembrane space TIM23 complex
50 23
Mitochondrial inner membrane
23 17
17 44
44 Matrix
Matrix, inner and intermembrane polypeptides
Mutant huntingtin protein PolyQ
Mutant huntingtin peptide N
N Figure 1 Mutant huntingtin inhibits mitochondrial protein import. Shown are the cytosolic compartment, mitochondrial inner and outer membranes, intermembrane space, and mitochondrial matrix. Mitochondrial preproteins translocate through the TOM complex in the outer membrane. Matrix and many inner membrane proteins are recognized by TIM23 subunits and translocate through the TIM23 complex. Mutant huntingtin full-length protein or mutant huntingtin N-terminal peptide, both containing long polyglutamine (polyQ) repeat sequences, bind the portion of the TIM23 complex consisting of Tim23, Tim17a and Tim50 (magenta), inhibiting protein import.
volume 17 | number 6 | June 2014 nature neuroscience
Marina Corral Spence/Nature Publishing Group
npg
© 2014 Nature America, Inc. All rights reserved.
news and views
npg
© 2014 Nature America, Inc. All rights reserved.
news and views proteins destined for the matrix or inner membrane, including many members of complexes that carry out intermediary metabolism and respiration. If mutant huntingtin is stuck to the TIM23 complex, it might disrupt protein import into the matrix or inner membrane, severely hampering metabolism in cells such as neurons that are highly energy dependent. Using an established assay, the investigators found that mutant huntingtin N-terminal peptides inhibited protein import into mitochondria isolated from mouse brain in a manner dependent on the concentration and length of the huntingtin polyglutamine sequence. The authors then tested protein import in vivo in a mouse model of Huntington’s disease. The R6/2 mouse expresses an N-terminal fragment from a human HTT gene and has become an important model because its disease progression mimics the motor abnormalities, metabolic disarray and shortened life span found in human patients10. Perhaps suspecting the importance of synaptic dysfunction in neurodegeneration, the authors isolated highly purified synaptosomal mitochondria (those residing near presynaptic terminals) in addition to non-synaptosomal mitochondria from the mutant mice. Synaptosomal mitochondria are a more select group. They are taken only from cells with synapses—that is, neurons— and may also function differently from mitochondria isolated from the soma because these specialized organelles buffer calcium and provide energy for synaptic activity. Because there are many demands placed on synaptic mitochondria, the authors guessed that their function might become more easily compromised by the mutant protein. Notably, the authors found that the synaptic mitochondria exhibited markedly compromised protein import even at presymptomatic and early disease stages (when the mice were 3 weeks old), whereas nonsynaptic mitochondria were protected from a decline in protein import even as late as 10–11 weeks of life. Mitochondria from liver also showed no deficit in protein import until much later in the disease, suggesting that synaptic mitochondria are particularly sensitive to these mutations. Even mitochondria isolated from embryonic neurons made from the R6/2 mouse before birth displayed the protein
import defect, suggesting that this deficit is one of the earliest manifestations of disease in mutant mouse synapses. If protein import into mitochondria is needed for normal mitochondrial respiration, then the delay in protein import should precede the onset of any respiratory chain dysfunction. The authors found that there were no deficits in respiratory function in mitochondria isolated from presymptomatic animals, even though these mitochondria showed signs of severe dysfunction in protein import, whereas respiratory function became quite compromised later in the disease. Of particular note, although the deficits in respiratory chain function have previously been attributed to high levels of reactive oxygen species (ROS) causing protein and DNA damage to vulnerable mitochondria, and this may still be true, Yano et al.4 show that sublethal doses of ROS, although further damaging protein import in mitochondria from mutant huntingtin–expressing cells, do not damage the protein import pathway in normal mitochondria, ruling out ROS as the primary cause of the protein import deficit. Is the change in protein import itself sufficient to produce neuronal dysfunction and death? To address this, the authors bypassed the mutant huntingtin protein to study the effect of depletion of the general import machinery directly. They first depleted the outer membrane translocator protein TOM40. They noted that cells died by apoptotic and non-apoptotic mechanisms soon after halting protein import. Next, the authors depleted the actual target of mutant huntingtin, the inner membrane translocator component TIM23. In live neurons, they were able to watch the loss of fluorescence of a voltage-dependent mitochondrial membrane potential indicator and the accumulation of a fluorescent cell death marker as the levels of TIM23 declined. Neuronal death occurred rapidly after loss of mitochondrial membrane potential, within 1–2 days of the onset of TIM23 depletion. The authors also found that increasing TIM23 in mutant huntingtin–expressing neurons produced an improvement in metabolic health and in neuronal survival. These striking findings—that reduced mitochondrial protein import caused by mutant huntingtin leads to metabolic dysfunction and death of neurons—raise
nature neuroscience volume 17 | number 6 | June 2014
many questions. Why do some neurons die and others survive if mutant huntingtin is expressed in all neurons? Is the amount of mutant huntingtin bound to the import machinery different in different cell populations, and, if so, how does this come about? And what about the intriguing finding that synaptic mitochondrial protein import is more affected by mutant huntingtin than that of somatic mitochondria? Is this discrepancy caused by a different level of activity of these mitochondria, increasing their vulnerability through an increase in reactive oxygen species? In turn, do these defective synaptic mitochondria cause some of the motor and cognitive symptoms of patients or animals found preceding cell death? If the protein import deficit causes neuronal death so rapidly, why do most patients develop neuronal loss and clinical disease late in life, and why is there often no neuronal loss in animal models? Even though there are still so many unanswered questions, the findings by Yano et al. 4 that mutant huntingtin causes loss of mitochondrial protein import, resulting in a decline in metabolism and respiration, certainly provides an influx of ideas to the neuro degeneration story. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Harris, J.J., Jolivet, R. & Attwell, D. Neuron 75, 762–777 (2012). 2. Petrozzi, L., Ricci, G., Giglioli, N.J., Siciliano, G. & Mancuso, M. Biosci. Rep. 27, 87–104 (2007). 3. Chaturvedi, R.K. & Beal, M.F. Ann. NY Acad. Sci. 1147, 395–412 (2008). 4. Yano, H. et al. Nat. Neurosci. 17, 822–831 (2014). 5. Kumar, A., Vaish, M. & Ratan, R.R. Drug Discov. Today doi:10.1016/j.drudis.2014.03.016 (21 March 2014). 6. Oliveira, J.M. J. Neurochem. 114, 1–12 (2010). 7. Kuemmerle, S. et al. Ann. Neurol. 46, 842–849 (1999). 8. Brustovetsky, N. et al. J. Neurosci. 23, 4858–4867 (2003). 9. Grünewald, T. & Beal, M.F. Ann. NY Acad. Sci. 893, 203–213 (1999). 10. Crook, Z.R. & Housman, D. Neuron 69, 423–435 (2011). 11. Reddy, P.H. & Shirendeb, U.P. Biochim. Biophys. Acta 1822, 101–110 (2012). 12. Anderson, S. et al. Nature 290, 457–465 (1981). 13. Calvo, S.E. & Mootha, V.K. Annu. Rev. Genomics Hum. Genet. 11, 25–44 (2010). 14. Rehling, P., Wiedemann, N., Pfanner, N. & Truscott, K.N. Crit. Rev. Biochem. Mol. Biol. 36, 291–336 (2001).
749
