An N-terminal Nuclear Export Signal Regulates Trafficking and Aggregation of Huntingtin (Htt) Protein Exon 1*

Background: Trafficking of huntingtin (Htt) fragments influences its toxicity. Results: A leucine-rich NES lies within the first 17 amino acids (N17) of Htt that controls subcellular localization and aggregation. Conclusion: The NES functions in cis and regulates the aggregation of Htt. Significance: This helps explain the mechanism of subcellular trafficking and aggregation of Htt fragments and may help elucidate molecular mechanisms of Htt toxicity. Huntington disease is a dominantly inherited neurodegenerative condition caused by polyglutamine expansion in the N terminus of the huntingtin protein (Htt). The first 17 amino acids (N17) of Htt play a key role in regulating its toxicity and aggregation. Both nuclear export and cytoplasm retention functions have been ascribed to N17. We have determined that N17 acts as a nuclear export sequence (NES) within Htt exon and when fused to yellow fluorescent protein. We have defined amino acids within N17 that constitute the nuclear export sequence (NES). Mutation of any of the conserved residues increases nuclear accumulation of Htt exon 1. Nuclear export of Htt is sensitive to leptomycin B and is reduced by knockdown of exportin 1. In HEK293 cells, NES mutations decrease overall Htt aggregation but increase the fraction of cells with nuclear inclusions. In primary cultured neurons, NES mutations increase nuclear accumulation and increase overall aggregation. This work defines a bona fide nuclear export sequence within N17 and links it to effects on protein aggregation. This may help explain the important role of N17 in controlling Htt toxicity.

Huntington disease is a dominantly inherited neurodegenerative condition caused by polyglutamine expansion in the N terminus of the huntingtin protein (Htt). The first 17 amino acids (N17) of Htt play a key role in regulating its toxicity and aggregation. Both nuclear export and cytoplasm retention functions have been ascribed to N17. We have determined that N17 acts as a nuclear export sequence (NES) within Htt exon and when fused to yellow fluorescent protein. We have defined amino acids within N17 that constitute the nuclear export sequence (NES). Mutation of any of the conserved residues increases nuclear accumulation of Htt exon 1. Nuclear export of Htt is sensitive to leptomycin B and is reduced by knockdown of exportin 1. In HEK293 cells, NES mutations decrease overall Htt aggregation but increase the fraction of cells with nuclear inclusions. In primary cultured neurons, NES mutations increase nuclear accumulation and increase overall aggregation. This work defines a bona fide nuclear export sequence within N17 and links it to effects on protein aggregation. This may help explain the important role of N17 in controlling Htt toxicity.
Huntington disease (HD) 2 is an autosomal dominant neurodegenerative condition caused by expansion of the polyglutamine tract in the amino (N)-terminal region of the huntingtin protein (Htt) (1). Polyglutamine tract length determines Htt propensity for aggregation and toxicity in vitro, and age of onset in patients (2) (3). Htt is subjected to apparent proteolytic cleavage in vivo. This generates a spectrum of N-terminal fragments that are more aggregation-prone and toxic than the full-length protein (4 -7). Htt fragments accumulate in the nucleus in vivo, constituting an important pathological feature of HD (8,9). Htt exon 1, which includes the first 17 amino acids (N17), the polyglutamine tract, and a polyproline domain, is one of the most widely studied Htt-derived peptides. In addition to the polyglutamine tract, N17 plays an important role in Htt aggregation, clearance, and toxicity (10 -12). Several sites of posttranslational modification, including phosphorylation, ubiquitination, and sumoylation (11,13,14), have been identified in the N17 region and could potentially affect Htt pathogenesis. A recent study identified phosphoserine residues at Ser-13 and Ser-16 as key inhibitors of toxicity of the full-length Htt protein, and of the aggregation potential of a peptide fragment (10). It has also been reported that hyperphosphorylation of Ser-13 and Ser-16 promotes relocation of Htt from the cytoplasm to the nucleus (11,12,15). This region is thus very important in regulating Htt aggregation, subcellular localization, and toxicity.
Htt fragments have been associated with multiple organelles and are found predominantly in the cytoplasm (9,16). Previous work has identified a nuclear export sequence (NES) in the C terminus of Htt outside the exon 1 region (17). Additionally, deletion of N17 results in nuclear accumulation of Htt fragments, showing the importance of N17 in cytoplasmic localization (20). Work by others has suggested that N17 contains a temperature-sensitive cytoplasm retention signal that mediates association with the endoplasmic reticulum and mitochondria (13,18,19). Given the critical role of nuclear localization in Htt toxicity (20,21), we sought to define the mechanism for the effects of N17 on the trafficking of N-terminal Htt fragments. We have defined an NES motif in the N17 region and have studied its effects on subcellular localization and aggregation of Htt exon 1 in non-neural cells and primary cultured neurons.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The Htt exon 1 (Q25)-CFP/YFP and Htt exon 1 (Q72)-CFP/YFP expression plasmids (termed Htt(Qn)-CFP/YFP here) have been described previously (22) and consist of Htt exon 1 sequence fused in frame to coding sequence for cyan or yellow fluorescent proteins. The chimeric construct N17-YFP was generated by fusing the N17 sequence of Htt to the N terminus of YFP. All point mutants in the N17 region were generated using the QuikChange TM Site-Directed Mutagenesis Kit from Stratagene. For the preparation of lentivirus, Htt(Q25/Q72)-CFP/YFP and relevant mutants were inserted between the NheI and AscI sites of the parental lentiviral vector FC-FM5, which was derived from the lentiviral vector FCIV (a kind gift from the Milbrandt laboratory at Washington University in St. Louis) by deleting the IRES-Venus sequence.
Cell Culture and Transfection-HEK293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum (FBS) and transfected with Lipofectamine TM reagent. For fluorescence resonance energy transfer (FRET) assays, 24 h after transfection, cells were passaged onto a 96-well black plate, cultured for 24 h, and then fixed with 4% paraformaldehyde. For microscopic imaging, 24 h after transfection cells were fixed with 4% paraformaldehyde, and the nucleus was counterstained with DAPI. Fluorescence images of cells were taken using a confocal image system from Zeiss. For the culture of primary neurons, the cortex of E18 mouse embryos was isolated and digested with 2 mg/ml papain and 0.1% DNase I. Neurons in neurobasal media containing serum-free B-27 and Glutamax were then seeded on culture plates or glass coverslips, precoated with 10 g/ml poly-D-lysine and 1.5 g/cm 2 laminin. Medium was changed once every 3 days until neurons were ready for use.

Biochemical Quantification of Cytoplasmic and Nuclear
Fractionation-The cytoplasmic and nuclear fractions of HEK293 cells expressing Htt(Q25)-YFP and relevant mutants were prepared using the NE-PER cytoplasmic and nuclear extraction kit (Thermo Scientific). Western blotting was performed with these cytoplasmic and nuclear samples using antibodies against GFP, HSP90, and histone H1.
Quantification of Subcellular Localization by High Content Microscopy-The relative amounts of Htt-YFP in the nucleus and cytoplasm were quantified using an IN Cell Analyzer 1000 microscope (GE Healthcare), using a 10ϫ objective. Analysis was performed using the Multi-Target Analysis Module of the IN Cell Analyzer 1000 Work station 3.7 analysis software. A "decision tree" was programmed to select only transfected cells expressing Htt-YFP or N17-YFP, based upon user-defined intensity threshold and location as a cell body associated with a nucleus (as defined by coincident DAPI stain), which were then classified according to the ratio of nuclear/cytoplasmic (not whole cell) intensities in the green channel. Within a single experiment, transfections were performed in duplicate. Approximately 500 -600 cells were analyzed per transfection, or about 1000 -1200 individual cells analyzed for each construct. Results represent at least three replicates of transfections performed on separate days. A, homology between Htt N17 and a leucine-rich NES motif (L, hydrophobic amino acids, often leucine; X, random amino acids) is shown. B, single mutations of leucine 4, 7, 14 and phenylalanine 11 to serine or glycine, which are not hydrophobic, caused nuclear accumulation of Htt(Q25)-YFP. Mutations of the non-NES residues threonine 3, lysine 9, and serine 13 and 16 had no effect on the localization. DNA was stained with DAPI to visualize the nucleus. C, high content microscopy was used for quantification of the relative amount of nuclear versus cytoplasmic Htt(Q25)-YFP for each mutant. ANOVA versus WT as the control was performed (*, p Ͻ 0.01). D, Western blot of cytoplasmic (C) and nuclear (N) fractions of HEK293 cells expressing Htt(Q25)-YFP are shown using antibody against GFP. HSP90 and histone H1 were used as markers for cytoplasmic and nuclear fractions, respectively. Error bars represent the S.E.

Measuring Aggregation by FRET on a Fluorescence Plate
Reader-The total aggregation for every construct was measured by an assay based on FRET between the enhanced cyan and yellow fluorescent protein tags on Htt exon 1, which has been described previously (22). Every experiment included mock-transfected cells and cells transfected only with Htt-CFP or Htt-YFP. Prior to calculations all values were backgroundsubtracted. From a Htt-YFP well, the amount of YFP activation from 435-nm excitation was determined and divided by the amount of signal derived from excitation at 485 nm (termed X, usually ϳ0.05-0.08). The measured FRET meas value of the sample was determined by recording the signal at 435 nm (ex) and 527 nm (em) . The calculated FRET calc ϭ (FRET meas Ϫ sample (485ex/528em) * X)/sample (435ex/485em) . The FRET calculations provide a reproducible measure of intracellular aggregation and take into account the relative amounts of both donor and acceptor proteins. The relative FRET of all constructs with N17 point mutations was calculated by normalizing to the FRET of Htt(Q72)-CFP/YFP, which was arbitrarily given a value of 1.0. We controlled for total protein expression (using fluorescence intensity of CFP and YFP) by separately titrating transfected DNA to produce the same levels of donor and acceptor between experiments. This was done to ensure that changes in overall FRET were not due to simple up-or down-regulation of protein levels in the mutants. Additionally, our paradigm utilizes a ratiometric parameter that also controls independently for the level of FRET to donor signal. The FRET assay in mouse primary neurons was performed by co-infecting neurons with lentivirus carrying either Htt(Q72)-CFP or Htt(Q72)-YFP, after first independently confirming the expression efficiencies of each virus to ensure that for each N17 mutant similar expression levels were compared between each experiment.
FRET Measurements by Microscopy with Photobleaching-FRET donor-only expressing cells were transfected with 75 ng of Htt(Q72)-CFP and 225 ng of empty vector plasmid (pcDNA3.1). FRET acceptor-only expressing cells were transfected with 225 ng of Htt(Q72)-pYFP and 75 ng of empty vector plasmid (pcDNA 3.1). Cells expressing both donor and acceptor were transfected with 75 ng of Htt(Q72)-CFP and 225 ng of Htt(Q72)-pYFP. Twenty-four hours after transfection, cells were replated on Ibidi -slides (Ibidi GmbH) for imaging by microscopy. Cells were then cultured for an additional 24 h prior to fixation with 4% paraformaldehyde and microscopy analysis.
All images were obtained using a C-Apochromat 40X/1.2 numerical aperture lens (Carl Zeiss Advanced Imaging Microscopy, Jena, Germany). Digital images were acquired using a Zeiss LSM510 Meta NLO multiphoton/confocal laser-scanning microscope system on the Zeiss Axiovert 200M. Channels used for imaging were as follows. The donor CFP was stimulated using a 458-nm argon laser, and fluorescence was collected with a 480 -520-nm band pass filter; the acceptor YFP was stimulated using a 514-nm argon laser, and fluorescence was collected with a long pass 560-nm filter (23,24). The FRET channel was collected by stimulating with the 458-nm argon laser and collecting fluorescence with the long pass 560-nm filter. To create an image in which the intensity reflected an estimate of FRET efficiency, on a pixel-by-pixel basis the value of the initial CFP image was subtracted from the final CFP image obtained after photobleaching, and this difference was multiplied by 100 and divided by the final CFP image intensity, 100 Ϫ (CFP final Ϫ CFP initial )/CFP final . Proper adjustments were made for partial acceptor photobleaching using acceptor YFPonly expressing cells and accidental bleaching of donor using donor CFP-only expressing cells. Image arithmetic and grayscale-to-color image conversion were done using National Institutes of Health ImageJ 1.44 software.
Counting Nuclear Inclusions-Nuclear inclusion formation for all constructs was measured by calculating the ratio of cells with nuclear aggregation to all aggregation-positive cells in 10 random microscopic fields at 20ϫ power. Inclusions that completely overlapped with DAPI stain were considered "nuclear inclusion-positive." These experiments were repeated in three independent biological replicates. Approximately ϳ100 cells were identified in each replicate, or ϳ300 cells for each experimental condition.
Inactivation of Exportin 1 by siRNA Knockdown or Leptomycin B-HEK293 cells were co-transfected with Htt(Q25)-YFP or N17-YFP, and three individual siRNAs against exportin 1, or scrambled siRNAs. 48 h after transfection, exportin 1 expres-FIGURE 2. N17 exhibits nuclear export activity in cis. A, HEK293 cells were transfected with constructs of YFP and N17-YFP, in which N17 was fused in-frame to the N terminus of YFP. YFP located mainly in the nucleus, whereas N17-YFP had increased cytoplasm localization, and mutations of L7S and F11G blocked this effect. B, high content microscopy was used for quantification of the relative amount of nuclear versus cytoplasmic YFP and N17-YFP mutants. ANOVA versus WT as the control was performed (*, p Ͻ 0.01). Error bars represent the S.E. sion was examined by Western blotting using the exportin 1 antibody (BD Transduction Laboratories). Leptomycin B (LMB) was used as a selective inhibitor of exportin 1. Twentyfour hours after transfection, 10 ng/ml LMB dissolved in methanol was added to the medium and washed away after 2 h. Cells were then fixed with 4% paraformaldehyde and counterstained with DAPI, and the localization of Htt exon 1-YFP was examined by confocal imaging and high content microscopy.
Lentivirus-mediated Expression of Htt-YFP in Primary Neurons, Quantification of Subcellular Localization, and FRET Assays-HEK293T cells were co-transfected with recombinant FC-FM5, envelope plasmid VSVG, and packaging plasmid PSP (kind gifts from the Jeffrey Milbrandt laboratory at Washington University in St. Louis). Forty-eight hours after transfection, culture medium containing lentivirus was collected, concentrated, and added to the mouse primary cortical neurons to express unexpanded or expanded Htt-YFP and the N17 mutants. After fixation with 4% paraformaldehyde, neurons were counterstained with DAPI, and the localization of Htt-YFP was examined by confocal imaging. For quantification of subcellular localization, three separate transductions were performed on separate coverslips. Cells were analyzed using ImageJ 1.44 software, determining the ratio of nucleus to cytoplasm signal intensity from a sample of 100 -200 cells for each condition. FRET assays were performed in triplicate transductions, with four wells of a 96-well plate counted for each transduction and then averaged. It is estimated that ϳ1000 transduced neurons were recorded from each well for the FRET assays. In each case, virus was titrated to produce equivalent expression levels each for Htt(Q72)-CFP and Htt(Q72)-YFP.

RESULTS
A Potential NES Motif in the N17 Region-Htt N17 is a highly conserved leucine-rich sequence. A rough NES consensus (LX (1-3) LX (2)(3) LXL, where L are hydrophobic amino acids, mostly leucine, and X are random amino acids) has been recognized that is nonetheless highly protein context-dependent for function (25). Comparison with the N17 region in Htt revealed similarity in the conserved hydrophobic residues and the spacing pattern, despite slight differences in the spacing of the last two conserved hydrophobic residues (Fig. 1A). We hypothesized that leucines 4, 7, 14, and phenylalanine 11 in N17 constituted a fully functional NES for Htt fragments. To test this hypothesis, we used site-specific mutation of these hydrophobic amino acids within the context of an Htt exon 1-YFP fusion carrying an unexpanded polyglutamine (Glu-25) tract (Htt(Q25)-YFP). This predominantly locates in the cytoplasm of HEK293 cells (Fig. 1B). Each mutation within the putative NES increased nuclear accumulation of Htt(Q25)-YFP, whereas mutations of other amino acids in the N17 region, such as threonine 3, lysine 9, and serine 13 and 16, had no effect on the localization (Fig. 1B). The accumulation of Htt(Q25)-YFP with NES mutations was also confirmed using automated quantitative high content microscopy (Fig. 1C). Finally, subcellular fractionation via biochemistry confirmed nuclear accumulation of Htt(Q25)-YFP with NES mutations (Fig. 1D). To confirm that N17 contains a bona fide NES and does not require other components of the Htt protein, we fused this sequence in isolation to the yellow fluorescent protein (N17-YFP). As expected, YFP alone localized in both nucleus and cytoplasm, whereas N17-YFP was shifted predominantly to the cytoplasm ( Fig. 2A). As for Htt-YFP, L7S and F11G mutations each disrupted nuclear export of N17-YFP, as confirmed by imaging and high content microscopy (Fig. 2, A and B). Our findings indicated that N17 alone is sufficient to drive nuclear export.
Nuclear Export of Htt Depends on Exportin 1-Nuclear export of proteins by a leucine-rich NES is mediated by Crm1 in yeast and exportin 1 in mammalian cells (26,27). A ternary complex consisting of the NES cargo, Ran-GTP, and Crm1/ exportin 1 is formed in the nucleus and subsequently translocates out of the nucleus through the nuclear pore. The ternary complex then dissociates in the cytoplasm (28). LMB, a selective inhibitor of Crm1 and exportin 1, disrupts the ternary complex, causing nuclear accumulation of proteins with a functional NES motif. We tested whether nuclear export of Htt-YFP and N17-YFP was also dependent on exportin 1 using pharmacologic and genetic manipulation. First, we treated cells transiently expressing these constructs for 2 h with 10 ng/ml LMB, which has been used previously for HEK293 cells (29,30). This resulted in nuclear accumulation of Htt(Q25)-YFP and N17-YFP (Fig. 3, A and B). In addition, exportin 1 knockdown using three individual siRNAs also caused nuclear accumulation of Htt(Q25)-YFP and N17-YFP as measured by high content microscopy (Fig. 3, C-E). We observed no toxicity associated with knockdown, by morphology or cell counts. Thus, trafficking of both Htt-YFP and N17-YFP is mediated by exportin 1.
N17 Is Functional in Expanded Htt-A prior report has indicated that Htt associates with nuclear pore protein translocation region (Tpr) in a polyglutamine-dependent fashion, in which expanded glutamine tracts disrupt this interaction and putatively promote nuclear accumulation (31). We tested whether expanded polyglutamine tracts would alter nuclear versus cytoplasm accumulation of Htt(Q72)-YFP and whether nuclear export of expanded Htt would be affected by mutations in N17. To begin, we only analyzed cells with diffusely distributed Htt-YFP protein, because it is not easy to measure differences in trafficking in the setting of inclusion formation. We transfected HEK293 cells with Htt(Q72)-YFP and selected cells with diffuse protein distribution for analysis using high content microscopy. We observed a ratio of nuclear to cytoplasm distribution similar to Htt(Q25)-YFP, and point mutations within N17 had identical effects as well (Fig. 4).
N17 Modulates Htt Aggregation-HD brain accumulates insoluble Htt inclusions, especially in the nucleus, which is enriched for a variety of short Htt fragments (8,9,32). In addition to polyglutamine tract length, flanking sequence elements such as N17 also modulate Htt structure and aggregation (33,34) and full-length toxicity in vivo (10). We observed that point mutations within N17 that disrupt NES function produced a distinct morphology of the nuclear inclusions, with a single large inclusion surrounded by multiple, small puncta. These were easily distinguished from the typical large, round shape of intracellular inclusions of Htt(Q72)-YFP (Fig. 5A).
Next, we used FRET recorded from a fluorescence plate reader to test the effect of point mutations within N17. We have previously established this assay to measure intracellular polyglutamine protein aggregation and to record subtle changes induced by chemical and genetic modifiers (22,(35)(36)(37). In our experiments, a fixed ratio of Htt-CFP:Htt-YFP fusions (1:3) was expressed in a cell population, and FRET was measured using spectroscopic methods and acceptor photobleaching microscopy. When transfected into HEK293 cells, point mutations in the N17 NES strongly reduced the overall aggregation propensity of Htt(Q72)-CFP/YFP based on spectroscopic FRET measured on a fluorescence plate reader (Fig. 5B). However, upon direct inspection, based on manual counting of hundreds of cells in multiple fields, the percentage of cells with nuclear inclusions was strongly increased (Fig. 5C). We confirmed that differences in apparent FRET efficiency were not due to struc- . The NES is functional in the context of expanded polyglutamine. A, single mutations of leucine 4, 7, 14 and phenylalanine 11 to serine or glycine resulted in nuclear accumulation of Htt(Q72)-YFP. Mutation of serine 13 and 16 to alanine had no effect, whereas mutation to aspartate caused mild nuclear accumulation. Cytoplasmic localization was not changed for mutations of non-NES residues, including threonine 3 and lysine 9. B, high content microscopy was used for quantification of the relative amount of nuclear versus cytoplasmic Htt(Q72)-YFP for the mutants. ANOVA versus WT as the control was performed (*, p Ͻ 0.01). Error bars represent the S.E. tural differences in the aggregates by photobleaching inclusions formed by the various N17 mutants. This indicated that inclusions formed by the various mutants had equivalent FRET efficiencies (Fig. 5, D and E), strongly suggesting that differences in overall FRET derived from changes in intracellular aggregation.
We next tested the effects of N17 mutations in primary mouse cortical neurons. We first transduced primary neurons with lentivirus to express Htt(Q25)-YFP or Htt(Q72)-YFP. Both unexpanded and expanded forms with wild-type N17 were excluded from the nucleus, based on evaluation of Ͼ300 transduced cells for each condition (Fig. 6, A, B, and D). We separately evaluated cells with evident inclusions. In contrast to HEK293 cells, we observed that inclusions from wild-type Htt(Q72)-YFP formed almost exclusively in the nucleus (Fig.  6C). Mutation of the NES (L7S and F11G) caused nuclear accumulation, whereas K9A had no effect (Fig. 6C). Due to the small size of the neuronal cell body, we were not able to quantify the nucleus:cytoplasm ratio using high content imaging. However, we were able to study a small selection of representative cells using ImageJ analysis, which confirmed our findings based on visual inspection (Fig. 6D). The plate reader-based FRET assay allowed analysis of aggregation rates of WT Htt(Q72)-YFP and the N17 mutants in thousands of transduced neurons. In contrast to HEK293 cells, mutations of L7S and F11G substantially increased overall aggregation based on FRET (Fig. 6E). Thus, the NES of N17 functions in primary neurons and its disruption by point mutations strongly increase overall Htt(Q72)-YFP aggregation within the nucleus.

DISCUSSION
The N17 region is highly conserved and plays an important role in regulating toxicity of the expanded protein. There remains considerable debate about the molecular mechanisms that govern the subcellular localization of Htt fragments. Here we have found that N17 contains a bona fide NES, consisting of four hydrophobic amino acids, leucine 4, 7, 14, and phenylalanine 11. The NES functions in Htt with both unexpanded and expanded glutamine tracts and when fused to a heterologous protein (YFP). The nuclear export of Htt(Q25)-YFP and N17-YFP depends on the classic pathway for a leucine-rich NES, as inactivation of exportin 1 by LMB and siRNA knockdown each increased nuclear accumulation of these proteins. N17 NES activity also regulates the degree of nuclear inclusion formation. In primary neurons, where Htt inclusion formation is predominantly nuclear, disruption of the NES strongly increased overall Htt-YFP aggregation. In non-neural cells, where most inclusion formation occurs in the cytosol, disruption of the NES increased nuclear inclusion formation, but caused an overall reduction in total intracellular aggregation.
This work helps clarify prior studies by various groups with contrasting findings regarding this domain (11,13,17,19,31,32,38). Truant and colleagues have demonstrated that N17 Relative FRET for all constructs was calculated by normalizing to WT, which is presented as a mean of 1. All comparisons controlled for total protein expression of each mutant based on CFP and YFP fluorescence, so that cells with similar expression levels were analyzed in each assay. ANOVA versus WT as the control was performed (*, p Ͻ 0.01). C, NES mutants increase nuclear inclusion frequency. Nuclear inclusion formation for indicated constructs was measured by calculating the ratio of cells with nuclear inclusions to all inclusion-positive cells for 10 random microscopic fields (ϳ100 cells) in three biological replicates. Inclusions that completely overlapped with DAPI stain were considered "nuclear inclusion-positive." ANOVA versus WT as the control was performed (*, p Ͻ 0.01). Error bars represent the S.E. D, example of acceptor photobleaching of a WT Htt(Q72)-CFP/YFP inclusion to determine FRET efficiency is shown. Acceptor (YFP) signal in an inclusion is strongly reduced following photobleaching. This increases donor signal and allows a calculation of FRET efficiency, as indicated by the histogram and image. E, analysis of FRET efficiency of ϳ20 inclusions from each of the three constructs shows no intrinsic difference in FRET efficiency based on the N17 mutations. This suggests that overall changes in FRET are due to differences in intracellular aggregation, rather than changes in donor/acceptor orientation within an aggregate. mediates cytoplasm retention by virtue of its amphipathic alpha-helical structure (19). This is consistent with the results of deletion of N17 on Htt distribution observed by Thompson and colleagues (18). However, the effects we observed with LMB and exportin1 knockdown argue against this as the only activity of N17. A cytoplasm retention signal would, in theory, not be affected by these interventions, whereas we observed nuclear accumulation of Htt in the setting of inhibition of nuclear export. Given its high degree of evolutionary conservation, beyond the four amino acid NES motif, it is highly plausible that N17 contains overlapping, but distinct functional regions and further, that its activities could be regulated. Li and colleagues previously described an N17-mediated interaction with Tpr (31), a protein associated with the nuclear pore that was subsequently implicated in the function of CRM-1, the yeast homologue of mammalian exportin 1 (39). Further, they observed an inhibition of nuclear export for expanded Htt. Although we did not find a polyglutamine-dependent effect on the NES function of N17 as they did, our results are consistent with the observation of an export activity linked to this region. Our work extends these findings by defining the precise amino acid motif that governs N17-mediated nuclear export and implicating a specific nuclear exporter in the process.
Htt aggregation is one of the most distinctive features of HD pathogenesis. We observed strong effects of N17 on total aggregation as a consequence of single amino acid substitutions, which is consistent with the effects that others have observed (18,19). Interestingly, whereas mutation of the NES reduced overall protein aggregation in HEK293 cells as detected by FRET, it strongly increased the fraction of cells with nuclear inclusion formation. In primary cultured neurons, by contrast, disruption of the NES activity resulted in a strong overall increase in aggregation that occurred almost exclusively within the nucleus. Thus, post-translational modifications of N17 could have particularly strong effects on neuronal toxicity of Htt. We anticipate that N17 could influence aggregation in several ways. First, point mutations in N17 could control intrinsic protein structure. This could occur through N17-mediated nucleation events (33) or "gatekeeping" effects on protein stability (34). We have previously observed that deletion of the N17 region reduces Htt overall aggregation in cells (40). Conclusions by our group and others that the N17 region "promotes" aggregation, however, are surely an oversimplification, especially because N17 mutations that strongly increase overall aggregation in primary neurons reduce aggregation in HEK293 cells. By regulating subcellular localization, and thus protein interactions and local concentration, and by directly altering protein conformation, N17 likely exerts a complex role in HD pathogenesis. How nuclear accumulation of Htt is regulated and whether this has clear effects on Htt toxicity in the context of the full-length protein remain to be determined. However, this work suggests several ways in which the N17 region could affect Htt toxicity by regulating subcellular localization and aggregation potential.
Acknowledgments-We thank Dr. Jeffrey Milbrandt and his laboratory for supplying lentivirus constructs and Dr. Najla Kfoury for production of primary cultured neurons. Jayne Marasa supervised high content microscopy in the Washington University High Throughput Screening Core, affiliated with the Molecular Imaging Center, the Siteman Cancer Center, Mallinckrodt Institute of Radiology, BRIGHT Institute, and the Washington University School of Medicine. FIGURE 6. Mutations in the NES change Htt exon 1 subcellular localization and inclusion formation in primary neurons. A lentiviral expression system was used to express unexpanded or expanded Htt-YFP in mouse primary neurons. Representative images are included. A, cytoplasmic Ͼ nuclear localization was observed for Htt(Q25)-YFP with WT N17 sequence. L7S and F11G mutations in N17 increased nuclear accumulation, whereas K9A had no effect on subcellular localization. B, similar distribution patterns for diffuse Htt(Q72)-YFP were observed. C, representative images show inclusions formed by Htt(Q72)-YFP with WT versus mutant N17 in mouse primary neurons. D, relative amounts of nuclear versus cytoplasmic Htt-YFP in cells are quantified. ImageJ was used to compare the nucleus:cytoplasm signal in ϳ100 -200 representative cells from each condition. E, NES mutants increased overall aggregation in neurons transduced with lentivirus expressing Htt(Q72)-CFP/YFP, as measured by spectroscopic FRET assay with a fluorescence plate reader. Relative FRET for all constructs was calculated by normalizing to WT, which is presented with a mean of 1. ANOVA versus WT as the control was performed (*, p Ͻ 0.01). Error bars represent the S.E.