- Email: [email protected]
Microsatellite Expansion Diseases: Repeat Toxicity Found in Translation
Neuron
Previews by enhancing the pooling of neuronal responses. In contrast, the study by Saleem et al. shows that when mice navigate virtual reality, i.e., a high-contrast signal, the amplitude of narrowband decreases, as expected from their previous results with drifting gratings. Rather than enhancing the transfer of visual information, Saleem et al. propose that narrowand wideband represent two different channels of information transfer for different types of visual information. Future studies must resolve how luminance signals are coded in parallel with contrast signals and how they influence each other along the visual pathway.
Berger, H. (1929). Arch. Psychiatr. Nervenkr. 87, 527–570. Cardin, J.A. (2016). J. Neurosci. 36, 10496–10504. Cardin, J.A., Palmer, L.A., and Contreras, D. (2005). J. Neurosci. 25, 5339–5350. Castelo-Branco, M., Neuenschwander, S., and Singer, W. (1998). J. Neurosci. 18, 6395–6410. Gray, C.M., and McCormick, D.A. (1996). Science 274, 109–113. Hubel, D.H. (1959). J. Physiol. 147, 226–238. Lee, A.M., Hoy, J.L., Bonci, A., Wilbrecht, L., Stryker, M.P., and Niell, C.M. (2014). Neuron 83, 455–466. Niell, C.M., and Stryker, M.P. (2010). Neuron 65, 472–479.
REFERENCES
Nusbaum, M.P., and Beenhakker, M.P. (2002). Nature 417, 343–350.
Adrian, E.D. (1950). Electroencephalogr. Clin. Neurophysiol. 2, 377–388.
Pinto, L., Goard, M.J., Estandian, D., Xu, M., Kwan, A.C., Lee, S.H., Harrison, T.C., Feng, G., and Dan, Y. (2013). Nat. Neurosci. 16, 1857–1863.
Saleem, A.B., Lien, A.D., Krumin, M., Haider, B., Roma´n, M., Aslı Aya, R., Reinhold, K., Busse, L., Carandini, M., and Harris, K.D. (2017). Neuron 93, this issue, 315–322. Steriade, M., Curro´ Dossi, R., and Contreras, D. (1993). Neuroscience 56, 1–9. Steriade, M., Contreras, D., Amzica, F., and Timofeev, I. (1996). J. Neurosci. 16, 2788–2808. Storchi, R., Bedford, R.A., Martial, F.P., Allen, A.E., Wynne, J., Montemurro, M.A., Petersen, R.S., and Lucas, R.J. (2017). Neuron 93, this issue, 299–307. Traub, R.D., Jefferys, J.G., and Whittington, M.A. (1999). Fast Oscillations in Cortical Circuits (MIT Press). Vinck, M., Batista-Brito, R., Knoblich, U., and Cardin, J.A. (2015). Neuron 86, 740–754. Welle, C.G., and Contreras, D. (2016). J. Neurophysiol. 115, 1821–1835. Wurtz, R.H. (1969). J. Neurophysiol. 32, 727–742.
Microsatellite Expansion Diseases: Repeat Toxicity Found in Translation Fen-Biao Gao1,* and Joel D. Richter2 1Department
of Neurology in Molecular Medicine University of Massachusetts Medical School, Worcester, MA 01605, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.neuron.2017.01.001 2Program
Sellier et al. (2017) show that translation of expanded CGG repeats in fragile X-associated tremor/ataxia syndrome is initiated at an upstream ACG near-cognate start codon. The resulting polyglycine-containing protein, but not repeat RNA, is pathogenic by disrupting the nuclear lamina. Microsatellites, or short repeat sequences of 3–6 nt, can be expanded up to thousands of times in protein-coding or noncoding regions of diverse genes and cause a wide array of neurodevelopmental and neurodegenerative disorders. The molecular pathology of these disorders includes loss of host gene function as in fragile X syndrome, production of toxic polypeptides as in Huntington’s disease, and sequestration of RNA binding proteins by repeat-containing RNA as in myotonic dystrophies. In 2011, Laura Ranum and colleagues described a new molecular mechanism
by which repeat expansions may be expressed and cause neurologic diseases: repeat-associated non-AUG (RAN) translation. In this process, protein synthesis is initiated not at the canonical AUG start codon, but instead directly on the expanded repeats and in all reading frames (Zu et al., 2011, 2013). These findings triggered a flurry of studies of several repeat expansion disorders to decipher whether these RAN translation products are toxic and, if so, how RAN translation is regulated at the molecular level. One of these repeat expansion disorders is fragile X-associated tremor/ataxia
syndrome (FXTAS). Its characteristics, which are generally late onset, include intention tremor, ataxia, Parkinson’s disease-like symptoms, and other clinical manifestations (Hagerman and Hagerman, 2016). The disorder is caused by CGG repeats that are expanded 55–200 times in the 50 UTR of FMR1 mRNA, which is referred to as the fragile X pre-mutation, while a full expansion of 200 repeats or more leads to transcriptional silencing of FMR1 and fragile X syndrome. In FXTAS, FMR1 mRNA containing a CGG expansion is expressed at an aberrantly high level, suggesting that it promotes
Neuron 93, January 18, 2017 ª 2017 Elsevier Inc. 249
Neuron
Previews ACG (CGG)99
(CGG)99
ATG
5’ UTR FMR1
GFP
ATG
5’ UTR FMR1
GFP
Translation
mG 7
ACG (CGG)99
LAP2β
AAAA
FMRpolyG N-terminus Glycine C-terminus
m7G
(CGG)99
AAAA
No Translation No RNA Toxicity
Figure 1. Proposed Pathogenic Mechanism in Mouse Models of Fragile X-Associated Tremor/Ataxia Syndrome
pathogenicity, although how it might do so is unclear. According to one school of thought, expanded CGG repeats form nuclear RNA foci that sequester several RNA-binding proteins, a mechanism reminiscent of that in myotonic dystrophies (e.g., Sellier et al., 2010). On the other hand, a polyglycinecontaining protein (FMRpolyG) thought to be produced through RAN translation accumulates in FXTAS patient brains and contributes to CGG repeat toxicity in a Drosophila model (Todd et al., 2013). To further differentiate the relative contributions of RNA-mediated toxicity and FMRpolyG to the pathogenesis of FXTAS, Sellier et al. (2017) established a novel mouse model in which 99 CGG repeats were expressed in the molecular context of the full-length 50 UTR of FMR1 followed by the GFP coding region (Figure 1). Using serial deletion constructs and proteomic analyses, the authors found that translation of mRNA containing the expanded CGG repeats is initiated at an upstream ACG nearcognate start codon embedded in a putative Kozak consensus sequence, an optimal context for initiation. The resulting FMRpolyG protein consists of an N terminus of 12 amino acids followed by polyglycine and a 42-amino acid C terminus. Interestingly, translation initiated at the ACG codon is independent of the CGG repeats, and expansion of over 70 repeats creates an upstream open reading frame (uORF) large enough to produce a stable FMRpolyG. Sellier et al. (2017) 250 Neuron 93, January 18, 2017
further demonstrated that FMRpolyG— but not FMRpolyA, which is predicted to be translated from a different frame of expanded CGG repeats—is detected by antibodies specific to both the N and C terminus in brain samples from FXTAS patients. These surprising results indicate that FMRpolyG is not initiated directly at the expanded CGG repeats and hence is not a product of RAN translation. Instead, translation is mediated through a conventional ribosome scanning mechanism: it begins at an upstream ACG start codon decoded by Met-tRNA as revealed by proteomic analysis, reads through 55–200 CGG codons, and terminates at a TAA stop codon in a different reading frame of the FMR1 coding region. Results from another cell-based assay system also suggest that translation of expanded CGG repeats commences upstream near-cognate start codons through a cap-dependent ribosome scanning mechanism (Kearse et al., 2016). Because the ACG codon initiates translation in a different reading frame relative to FMR1, there is no ‘‘polyprotein’’ composed of both polyG and FMRP. Moreover, uORFs often control the translational efficacy of the downstream ORF as a response to external stimulation, frequently through the initiation factor eIF2a (Hinnebusch et al., 2016). Indeed, Sellier et al. (2017) show that the ACG-initiated CGG repeat uORF imparts translational regulation FMR1.
Mice expressing the full-length 50 UTR of FMR1 with 99 CGG repeats, either ubiquitously or in neurons and glia only, exhibited neurodegenerative phenotypes such as accumulation of FMRpolyG aggregates and increased brain inflammation, Purkinje cell loss, and locomotor deficits. To dissect whether these phenotypes were caused by RNA toxicity or FMRpolyG, Sellier et al. (2017) established a second mouse model lacking the ACG near-cognate start codon and surrounding 50 UTR sequence; as a consequence, only CGG repeat RNA, but not FMRpolyG, was synthesized, suggesting that the expanded CGG repeats in and of themselves are not competent for translation in vivo (Figure 1). These mice had no signs of neurodegeneration at the cellular or behavioral level, indicating that FMRpolyG, but not repeat RNA, is the primary driver of toxicity in FXTAS. These results add fresh experimental evidence to the debate regarding RNA versus protein toxicity in other repeat expansion diseases such as C9ORF72related ALS/FTD (frontotemporal dementia) (Gitler and Tsuiji, 2016). Interestingly, nuclear RNA foci were largely absent in mice expressing either the full-length or mutant 50 UTR of FMR1. This is in stark contrast to previous reports that expanded CGG repeats form RNA foci in transfected cells that sequester RNA-binding proteins (e.g., Sellier et al., 2010). The current study shows that the 50 UTR of FMR1 influences both the subcellular localization of expanded CGG repeats and their ability to form foci. This finding strengthens the notion that the molecular context of expanded repeat sequences is critically important for their toxicity (Tran et al., 2015). Indeed, Sellier et al. (2017) demonstrated that the 42-amino acid C terminus of FMRpolyG induces toxicity in both cultured cells and Drosophila models, while polyglycine itself is required for aggregate formation. In artificial constructs expressing expanded CGG repeats without the surrounding FMR1 50 UTR sequence, the toxicity of the C terminus of endogenous FMRpolyG could not be assessed, further highlighting the importance of studying cellular and animal models in which expanded repeats are expressed in their native molecular contexts.
Neuron
Previews To understand the mechanism by which FMRpolyG induces neurotoxicity, Sellier et al. (2017) looked for interacting proteins. One protein that co-precipitated with FMRpolyG was lamina-associated polypeptide 2 beta (LAP2b), and the two proteins co-localized in nuclear aggregates in transfected cells and in brain tissues of patients with FXTAS. Moreover, expression of FMRpolyG in HEK293 cells or primary rodent neurons altered nuclear organization. The interaction between these two proteins is mediated by the C terminus of FMRpolyG, underscoring the functional significance of the polypeptide translated from sequences downstream of expanded CGG repeats. Remarkably, overexpression of LAP2b prevented neuronal cell death induced by full-length FMRpolyG or its C terminus alone. The functional significance of other proteins that interact with FMRpolyG remains to be determined. These observations from overexpression studies seem to be directly relevant in patient neurons. Sellier et al. (2017) generated induced pluripotent stem cell lines from control subjects and three FXTAS patients with 84–99 CGG repeats and differentiated the lines into cortical neurons. These neurons had an agedependent increase in the number of FMRpolyG nuclear aggregates that were also positive for LAP2b, and the nuclear lamina was disorganized. Defects in the nuclear lamina (Freibaum et al., 2015) and nucleocytoplasmic transport are also found in C9ORF72-related ALS/ FTD (Zhang et al., 2016), suggesting a common pathogenic mechanism in different neurodegenerative diseases. In sum, the physical interaction between these two proteins seems to play a key role in the pathogenesis of FXTAS and may be an excellent molecular target for therapeutic intervention. The findings reported here by Sellier et al. (2017) raise important questions for
studies of other repeat expansion diseases as well, such as C9ORF72-related ALS/FTD. The GGGGCC (G4C2) repeat expansion in the noncoding region of C9ORF72 is the most common genetic cause of both ALS and FTD, and the dipeptide repeat (DPR) proteins that can be produced through RAN translation are toxic in various cellular and animal models (Gitler and Tsuiji, 2016). The C terminus of some DPR proteins, like that of FMRpolyG in FXTAS (Sellier et al., 2017), is translated from sequences downstream of the expanded G4C2 repeats in patient brains (Zu et al., 2013). However, its effect on the toxicity and subcellular localization of DPR proteins has not been investigated. Similarly, in many artificial constructs expressing expanded G4C2 or other repeats, an unnatural polypeptide of unknown function will be translated from the vector sequence downstream of the repeats. Thus, caution is needed to interpret the toxicity induced by these repeat constructs. The actual lengths of different DPR proteins in human patient brains are unknown. We also do not understand how RAN translation is regulated in C9ORF72-related ALS/FTD. Because G4C2 repeats are located in the first intron of C9ORF72, it is also unclear whether an unspliced C9ORF72 pre-mRNA or a spliced intron containing G4C2 repeats serves as the template for RAN translation. Sellier et al. (2017) noticed that an upstream CUG near-cognate start codon in a Kozak sequence is in-frame with G4C2 repeats encoding poly(GA). This observation raises the intriguing possibility that expanded G4C2 repeats form a long ORF with a start codon. If so, translation of at least some DPR proteins may be initiated at this start codon by conventional ribosome scanning and through the G4C2 repeats that are part of either a cap-containing mRNA with a uORF or a spliced intron. Irrespective of the extent
to which protein synthesis occurs directly on the endogenous expanded repeats independently of any start codons, the toxicity and disease relevance of abnormal proteins translated from various expanded repeats will be under intense investigation for years to come.
REFERENCES Freibaum, B.D., Lu, Y., Lopez-Gonzalez, R., Kim, N.C., Almeida, S., Lee, K.H., Badders, N., Valentine, M., Miller, B.L., Wong, P.C., et al. (2015). Nature 525, 129–133. Gitler, A.D., and Tsuiji, H. (2016). Brain Res. 1647, 19–29. Hagerman, R.J., and Hagerman, P. (2016). Nat. Rev. Neurol. 12, 403–412. Hinnebusch, A.G., Ivanov, I.P., and Sonenberg, N. (2016). Science 352, 1413–1416. Kearse, M.G., Green, K.M., Krans, A., Rodriguez, C.M., Linsalata, A.E., Goldstrohm, A.C., and Todd, P.K. (2016). Mol. Cell 62, 314–322. Sellier, C., Rau, F., Liu, Y., Tassone, F., Hukema, R.K., Gattoni, R., Schneider, A., Richard, S., Willemsen, R., Elliott, D.J., et al. (2010). EMBO J. 29, 1248–1261. Sellier, C., Buijsen, R.A., He, F., Natla, S., Jung, L., Tropel, P., Gaucherot, A., Jacobs, H., Meziane, H., Vincent, A., et al. (2017). Neuron 93, this issue, 331–347. Todd, P.K., Oh, S.Y., Krans, A., He, F., Sellier, C., Frazer, M., Renoux, A.J., Chen, K.C., Scaglione, K.M., Basrur, V., et al. (2013). Neuron 78, 440–455. Tran, H., Almeida, S., Moore, J., Gendron, T.F., Chalasani, U., Lu, Y., Du, X., Nickerson, J.A., Petrucelli, L., Weng, Z., and Gao, F.-B. (2015). Neuron 87, 1207–1214. Zhang, K., Grima, J.C., Rothstein, J.D., and Lloyd, T.E. (2016). Nucleus 7, 132–137. Zu, T., Gibbens, B., Doty, N.S., Gomes-Pereira, M., Huguet, A., Stone, M.D., Margolis, J., Peterson, M., Markowski, T.W., Ingram, M.A., et al. (2011). Proc. Natl. Acad. Sci. USA 108, 260–265. Zu, T., Liu, Y., Ban˜ez-Coronel, M., Reid, T., Pletnikova, O., Lewis, J., Miller, T.M., Harms, M.B., Falchook, A.E., Subramony, S.H., et al. (2013). Proc. Natl. Acad. Sci. USA 110, E4968–E4977.
Neuron 93, January 18, 2017 251
Previews by enhancing the pooling of neuronal responses. In contrast, the study by Saleem et al. shows that when mice navigate virtual reality, i.e., a high-contrast signal, the amplitude of narrowband decreases, as expected from their previous results with drifting gratings. Rather than enhancing the transfer of visual information, Saleem et al. propose that narrowand wideband represent two different channels of information transfer for different types of visual information. Future studies must resolve how luminance signals are coded in parallel with contrast signals and how they influence each other along the visual pathway.
Berger, H. (1929). Arch. Psychiatr. Nervenkr. 87, 527–570. Cardin, J.A. (2016). J. Neurosci. 36, 10496–10504. Cardin, J.A., Palmer, L.A., and Contreras, D. (2005). J. Neurosci. 25, 5339–5350. Castelo-Branco, M., Neuenschwander, S., and Singer, W. (1998). J. Neurosci. 18, 6395–6410. Gray, C.M., and McCormick, D.A. (1996). Science 274, 109–113. Hubel, D.H. (1959). J. Physiol. 147, 226–238. Lee, A.M., Hoy, J.L., Bonci, A., Wilbrecht, L., Stryker, M.P., and Niell, C.M. (2014). Neuron 83, 455–466. Niell, C.M., and Stryker, M.P. (2010). Neuron 65, 472–479.
REFERENCES
Nusbaum, M.P., and Beenhakker, M.P. (2002). Nature 417, 343–350.
Adrian, E.D. (1950). Electroencephalogr. Clin. Neurophysiol. 2, 377–388.
Pinto, L., Goard, M.J., Estandian, D., Xu, M., Kwan, A.C., Lee, S.H., Harrison, T.C., Feng, G., and Dan, Y. (2013). Nat. Neurosci. 16, 1857–1863.
Saleem, A.B., Lien, A.D., Krumin, M., Haider, B., Roma´n, M., Aslı Aya, R., Reinhold, K., Busse, L., Carandini, M., and Harris, K.D. (2017). Neuron 93, this issue, 315–322. Steriade, M., Curro´ Dossi, R., and Contreras, D. (1993). Neuroscience 56, 1–9. Steriade, M., Contreras, D., Amzica, F., and Timofeev, I. (1996). J. Neurosci. 16, 2788–2808. Storchi, R., Bedford, R.A., Martial, F.P., Allen, A.E., Wynne, J., Montemurro, M.A., Petersen, R.S., and Lucas, R.J. (2017). Neuron 93, this issue, 299–307. Traub, R.D., Jefferys, J.G., and Whittington, M.A. (1999). Fast Oscillations in Cortical Circuits (MIT Press). Vinck, M., Batista-Brito, R., Knoblich, U., and Cardin, J.A. (2015). Neuron 86, 740–754. Welle, C.G., and Contreras, D. (2016). J. Neurophysiol. 115, 1821–1835. Wurtz, R.H. (1969). J. Neurophysiol. 32, 727–742.
Microsatellite Expansion Diseases: Repeat Toxicity Found in Translation Fen-Biao Gao1,* and Joel D. Richter2 1Department
of Neurology in Molecular Medicine University of Massachusetts Medical School, Worcester, MA 01605, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.neuron.2017.01.001 2Program
Sellier et al. (2017) show that translation of expanded CGG repeats in fragile X-associated tremor/ataxia syndrome is initiated at an upstream ACG near-cognate start codon. The resulting polyglycine-containing protein, but not repeat RNA, is pathogenic by disrupting the nuclear lamina. Microsatellites, or short repeat sequences of 3–6 nt, can be expanded up to thousands of times in protein-coding or noncoding regions of diverse genes and cause a wide array of neurodevelopmental and neurodegenerative disorders. The molecular pathology of these disorders includes loss of host gene function as in fragile X syndrome, production of toxic polypeptides as in Huntington’s disease, and sequestration of RNA binding proteins by repeat-containing RNA as in myotonic dystrophies. In 2011, Laura Ranum and colleagues described a new molecular mechanism
by which repeat expansions may be expressed and cause neurologic diseases: repeat-associated non-AUG (RAN) translation. In this process, protein synthesis is initiated not at the canonical AUG start codon, but instead directly on the expanded repeats and in all reading frames (Zu et al., 2011, 2013). These findings triggered a flurry of studies of several repeat expansion disorders to decipher whether these RAN translation products are toxic and, if so, how RAN translation is regulated at the molecular level. One of these repeat expansion disorders is fragile X-associated tremor/ataxia
syndrome (FXTAS). Its characteristics, which are generally late onset, include intention tremor, ataxia, Parkinson’s disease-like symptoms, and other clinical manifestations (Hagerman and Hagerman, 2016). The disorder is caused by CGG repeats that are expanded 55–200 times in the 50 UTR of FMR1 mRNA, which is referred to as the fragile X pre-mutation, while a full expansion of 200 repeats or more leads to transcriptional silencing of FMR1 and fragile X syndrome. In FXTAS, FMR1 mRNA containing a CGG expansion is expressed at an aberrantly high level, suggesting that it promotes
Neuron 93, January 18, 2017 ª 2017 Elsevier Inc. 249
Neuron
Previews ACG (CGG)99
(CGG)99
ATG
5’ UTR FMR1
GFP
ATG
5’ UTR FMR1
GFP
Translation
mG 7
ACG (CGG)99
LAP2β
AAAA
FMRpolyG N-terminus Glycine C-terminus
m7G
(CGG)99
AAAA
No Translation No RNA Toxicity
Figure 1. Proposed Pathogenic Mechanism in Mouse Models of Fragile X-Associated Tremor/Ataxia Syndrome
pathogenicity, although how it might do so is unclear. According to one school of thought, expanded CGG repeats form nuclear RNA foci that sequester several RNA-binding proteins, a mechanism reminiscent of that in myotonic dystrophies (e.g., Sellier et al., 2010). On the other hand, a polyglycinecontaining protein (FMRpolyG) thought to be produced through RAN translation accumulates in FXTAS patient brains and contributes to CGG repeat toxicity in a Drosophila model (Todd et al., 2013). To further differentiate the relative contributions of RNA-mediated toxicity and FMRpolyG to the pathogenesis of FXTAS, Sellier et al. (2017) established a novel mouse model in which 99 CGG repeats were expressed in the molecular context of the full-length 50 UTR of FMR1 followed by the GFP coding region (Figure 1). Using serial deletion constructs and proteomic analyses, the authors found that translation of mRNA containing the expanded CGG repeats is initiated at an upstream ACG nearcognate start codon embedded in a putative Kozak consensus sequence, an optimal context for initiation. The resulting FMRpolyG protein consists of an N terminus of 12 amino acids followed by polyglycine and a 42-amino acid C terminus. Interestingly, translation initiated at the ACG codon is independent of the CGG repeats, and expansion of over 70 repeats creates an upstream open reading frame (uORF) large enough to produce a stable FMRpolyG. Sellier et al. (2017) 250 Neuron 93, January 18, 2017
further demonstrated that FMRpolyG— but not FMRpolyA, which is predicted to be translated from a different frame of expanded CGG repeats—is detected by antibodies specific to both the N and C terminus in brain samples from FXTAS patients. These surprising results indicate that FMRpolyG is not initiated directly at the expanded CGG repeats and hence is not a product of RAN translation. Instead, translation is mediated through a conventional ribosome scanning mechanism: it begins at an upstream ACG start codon decoded by Met-tRNA as revealed by proteomic analysis, reads through 55–200 CGG codons, and terminates at a TAA stop codon in a different reading frame of the FMR1 coding region. Results from another cell-based assay system also suggest that translation of expanded CGG repeats commences upstream near-cognate start codons through a cap-dependent ribosome scanning mechanism (Kearse et al., 2016). Because the ACG codon initiates translation in a different reading frame relative to FMR1, there is no ‘‘polyprotein’’ composed of both polyG and FMRP. Moreover, uORFs often control the translational efficacy of the downstream ORF as a response to external stimulation, frequently through the initiation factor eIF2a (Hinnebusch et al., 2016). Indeed, Sellier et al. (2017) show that the ACG-initiated CGG repeat uORF imparts translational regulation FMR1.
Mice expressing the full-length 50 UTR of FMR1 with 99 CGG repeats, either ubiquitously or in neurons and glia only, exhibited neurodegenerative phenotypes such as accumulation of FMRpolyG aggregates and increased brain inflammation, Purkinje cell loss, and locomotor deficits. To dissect whether these phenotypes were caused by RNA toxicity or FMRpolyG, Sellier et al. (2017) established a second mouse model lacking the ACG near-cognate start codon and surrounding 50 UTR sequence; as a consequence, only CGG repeat RNA, but not FMRpolyG, was synthesized, suggesting that the expanded CGG repeats in and of themselves are not competent for translation in vivo (Figure 1). These mice had no signs of neurodegeneration at the cellular or behavioral level, indicating that FMRpolyG, but not repeat RNA, is the primary driver of toxicity in FXTAS. These results add fresh experimental evidence to the debate regarding RNA versus protein toxicity in other repeat expansion diseases such as C9ORF72related ALS/FTD (frontotemporal dementia) (Gitler and Tsuiji, 2016). Interestingly, nuclear RNA foci were largely absent in mice expressing either the full-length or mutant 50 UTR of FMR1. This is in stark contrast to previous reports that expanded CGG repeats form RNA foci in transfected cells that sequester RNA-binding proteins (e.g., Sellier et al., 2010). The current study shows that the 50 UTR of FMR1 influences both the subcellular localization of expanded CGG repeats and their ability to form foci. This finding strengthens the notion that the molecular context of expanded repeat sequences is critically important for their toxicity (Tran et al., 2015). Indeed, Sellier et al. (2017) demonstrated that the 42-amino acid C terminus of FMRpolyG induces toxicity in both cultured cells and Drosophila models, while polyglycine itself is required for aggregate formation. In artificial constructs expressing expanded CGG repeats without the surrounding FMR1 50 UTR sequence, the toxicity of the C terminus of endogenous FMRpolyG could not be assessed, further highlighting the importance of studying cellular and animal models in which expanded repeats are expressed in their native molecular contexts.
Neuron
Previews To understand the mechanism by which FMRpolyG induces neurotoxicity, Sellier et al. (2017) looked for interacting proteins. One protein that co-precipitated with FMRpolyG was lamina-associated polypeptide 2 beta (LAP2b), and the two proteins co-localized in nuclear aggregates in transfected cells and in brain tissues of patients with FXTAS. Moreover, expression of FMRpolyG in HEK293 cells or primary rodent neurons altered nuclear organization. The interaction between these two proteins is mediated by the C terminus of FMRpolyG, underscoring the functional significance of the polypeptide translated from sequences downstream of expanded CGG repeats. Remarkably, overexpression of LAP2b prevented neuronal cell death induced by full-length FMRpolyG or its C terminus alone. The functional significance of other proteins that interact with FMRpolyG remains to be determined. These observations from overexpression studies seem to be directly relevant in patient neurons. Sellier et al. (2017) generated induced pluripotent stem cell lines from control subjects and three FXTAS patients with 84–99 CGG repeats and differentiated the lines into cortical neurons. These neurons had an agedependent increase in the number of FMRpolyG nuclear aggregates that were also positive for LAP2b, and the nuclear lamina was disorganized. Defects in the nuclear lamina (Freibaum et al., 2015) and nucleocytoplasmic transport are also found in C9ORF72-related ALS/ FTD (Zhang et al., 2016), suggesting a common pathogenic mechanism in different neurodegenerative diseases. In sum, the physical interaction between these two proteins seems to play a key role in the pathogenesis of FXTAS and may be an excellent molecular target for therapeutic intervention. The findings reported here by Sellier et al. (2017) raise important questions for
studies of other repeat expansion diseases as well, such as C9ORF72-related ALS/FTD. The GGGGCC (G4C2) repeat expansion in the noncoding region of C9ORF72 is the most common genetic cause of both ALS and FTD, and the dipeptide repeat (DPR) proteins that can be produced through RAN translation are toxic in various cellular and animal models (Gitler and Tsuiji, 2016). The C terminus of some DPR proteins, like that of FMRpolyG in FXTAS (Sellier et al., 2017), is translated from sequences downstream of the expanded G4C2 repeats in patient brains (Zu et al., 2013). However, its effect on the toxicity and subcellular localization of DPR proteins has not been investigated. Similarly, in many artificial constructs expressing expanded G4C2 or other repeats, an unnatural polypeptide of unknown function will be translated from the vector sequence downstream of the repeats. Thus, caution is needed to interpret the toxicity induced by these repeat constructs. The actual lengths of different DPR proteins in human patient brains are unknown. We also do not understand how RAN translation is regulated in C9ORF72-related ALS/FTD. Because G4C2 repeats are located in the first intron of C9ORF72, it is also unclear whether an unspliced C9ORF72 pre-mRNA or a spliced intron containing G4C2 repeats serves as the template for RAN translation. Sellier et al. (2017) noticed that an upstream CUG near-cognate start codon in a Kozak sequence is in-frame with G4C2 repeats encoding poly(GA). This observation raises the intriguing possibility that expanded G4C2 repeats form a long ORF with a start codon. If so, translation of at least some DPR proteins may be initiated at this start codon by conventional ribosome scanning and through the G4C2 repeats that are part of either a cap-containing mRNA with a uORF or a spliced intron. Irrespective of the extent
to which protein synthesis occurs directly on the endogenous expanded repeats independently of any start codons, the toxicity and disease relevance of abnormal proteins translated from various expanded repeats will be under intense investigation for years to come.
REFERENCES Freibaum, B.D., Lu, Y., Lopez-Gonzalez, R., Kim, N.C., Almeida, S., Lee, K.H., Badders, N., Valentine, M., Miller, B.L., Wong, P.C., et al. (2015). Nature 525, 129–133. Gitler, A.D., and Tsuiji, H. (2016). Brain Res. 1647, 19–29. Hagerman, R.J., and Hagerman, P. (2016). Nat. Rev. Neurol. 12, 403–412. Hinnebusch, A.G., Ivanov, I.P., and Sonenberg, N. (2016). Science 352, 1413–1416. Kearse, M.G., Green, K.M., Krans, A., Rodriguez, C.M., Linsalata, A.E., Goldstrohm, A.C., and Todd, P.K. (2016). Mol. Cell 62, 314–322. Sellier, C., Rau, F., Liu, Y., Tassone, F., Hukema, R.K., Gattoni, R., Schneider, A., Richard, S., Willemsen, R., Elliott, D.J., et al. (2010). EMBO J. 29, 1248–1261. Sellier, C., Buijsen, R.A., He, F., Natla, S., Jung, L., Tropel, P., Gaucherot, A., Jacobs, H., Meziane, H., Vincent, A., et al. (2017). Neuron 93, this issue, 331–347. Todd, P.K., Oh, S.Y., Krans, A., He, F., Sellier, C., Frazer, M., Renoux, A.J., Chen, K.C., Scaglione, K.M., Basrur, V., et al. (2013). Neuron 78, 440–455. Tran, H., Almeida, S., Moore, J., Gendron, T.F., Chalasani, U., Lu, Y., Du, X., Nickerson, J.A., Petrucelli, L., Weng, Z., and Gao, F.-B. (2015). Neuron 87, 1207–1214. Zhang, K., Grima, J.C., Rothstein, J.D., and Lloyd, T.E. (2016). Nucleus 7, 132–137. Zu, T., Gibbens, B., Doty, N.S., Gomes-Pereira, M., Huguet, A., Stone, M.D., Margolis, J., Peterson, M., Markowski, T.W., Ingram, M.A., et al. (2011). Proc. Natl. Acad. Sci. USA 108, 260–265. Zu, T., Liu, Y., Ban˜ez-Coronel, M., Reid, T., Pletnikova, O., Lewis, J., Miller, T.M., Harms, M.B., Falchook, A.E., Subramony, S.H., et al. (2013). Proc. Natl. Acad. Sci. USA 110, E4968–E4977.
Neuron 93, January 18, 2017 251