| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Ataxin-3 with an altered conformation that exposes the polyglutamine domain is associated with the nuclear matrix
Introduction
Results
Expression of ataxin-3 within the nucleus exposes the polyglutamine epitope
1C2 binding to intranuclear ataxin-3 does not correlate with proteolysis
Ataxin-3 with an altered conformation is associated with the nuclear matrix
Discussion
Materials And Methods
Expression constructs and transfection
Immunofluorescence and confocal microscopy
Cellular fractionation
Western blotting
In situ nuclear matrix preparation
Acknowledgements
References
Ataxin-3 with an altered conformation that exposes the polyglutamine domain is associated with the nuclear matrix
Received June 18, 1999; Revised and Accepted September 27, 1999
Spinocerebellar ataxia type-3 or Machado-Joseph disease (SCA3/MJD) is a member of the CAG/polyglutamine repeat disease family. In this family of disorders, a normally polymorphic CAG repeat becomes expanded, resulting in expression of an expanded polyglutamine domain in the disease gene product. Experimental models of polyglutamine disease implicate the nucleus in pathogenesis; however, the link between intranuclear expression of expanded polyglutamine and neuronal dysfunction remains unclear. Here we demonstrate that ataxin-3, the disease protein in SCA3/MJD, adopts a unique conformation when expressed within the nucleus of transfected cells. The monoclonal antibody 1C2 is known preferentially to bind expanded polyglutamine, but we find that it also binds a fragment of ataxin-3 containing a normal glutamine repeat. In addition, expression of ataxin-3 within the nucleus exposes the glutamine domain of the full-length non-pathological protein, allowing it to bind the monoclonal antibody 1C2. Fractionation and immunochemical experiments indicate that this novel conformation of intranuclear ataxin-3 is not due to proteolysis, suggesting instead that association with nuclear protein(s) alters the structure of full-length ataxin-3 which exposes the polyglutamine domain. This conformationally altered ataxin-3 is bound to the nuclear matrix. The pathological form of ataxin-3 with an expanded polyglutamine domain also associates with the nuclear matrix. These data suggest that an early event in the pathogenesis of SCA3/MJD may be an altered conformation of ataxin-3 within the nucleus that exposes the polyglutamine domain.
INTRODUCTION
Spinocerebellar ataxia type-3 or Machado-Joseph disease (SCA3/MJD) is a member of the CAG/polyglutamine repeat disease family. Other diseases within this family include Huntington's disease (HD), spinal and bulbar muscular atrophy (SBMA), dentatorubral pallidoluysian atrophy (DRPLA) and several other spinocerebellar ataxias (SCAs), including types 1, 2, 6 and 7 (1). The genetic basis underlying these disorders is a CAG trinucleotide repeat expansion in the protein-coding region, resulting in expression of an expanded polyglutamine domain in the otherwise unrelated disease proteins. Expanded polyglutamine confers a toxic gain of function on the disease protein, presumably through an increased propensity towards aggregation, altered protein interactions or both (2,3). Ataxin-3, the disease protein involved in SCA3/MJD, is a ubiquitously expressed protein of unknown function with a predicted molecular weight of 42 kDa (4-6). Normally it contains 12-40 glutamines near the C-terminus, and in disease the polyglutamine domain becomes expanded to 55-84 glutamines (7).
Accumulating experimental evidence in both transgenic animals and cell-based models of disease implicates the nucleus as a primary site of pathogenesis for several members of the polyglutamine disease family, including SCA3/MJD. Consistent with this, nuclear inclusions of pathological protein have been identified in affected neuronal populations in patients with SCA3/MJD as well as those with other polyglutamine diseases (8-11). The importance of the nuclear environment has been demonstrated clearly in a transgenic mouse model of SCA1 (12). Mice expressing mutant ataxin-1, the resident nuclear protein responsible for SCA1, develop severe ataxia, cerebellar degeneration and nuclear inclusions. However, when mutant ataxin-1 is engineered to remain in the cytoplasm through mutagenesis of its nuclear localization signal (NLS), the mice do not develop ataxia or pathological changes (13). In addition, in a neuronal model of HD, placing a nuclear export signal on a polyglutamine-containing fragment of the mutant huntingtin protein significantly reduces the toxicity of this fragment (14). Therefore, although different polyglutamine disease proteins may normally be located primarily in the nucleus or the cytoplasm, data are consistent with nuclear localization being a critical event leading to neuronal dysfunction for a number of these diseases.
The nature of the relationship between the nuclear environment and the toxic gain of function associated with the expanded polyglutamine domain remains a key unanswered question. Perhaps the expanded polyglutamine domain alters nuclear import or export of the disease protein. Alternatively, polyglutamine expansion could modify interactions with nuclear proteins, thus altering their function. Another possibility is that repeat expansion could result in a conformational change of the disease protein which, in the nucleus, alters susceptibility to proteolysis or exposes the polyglutamine domain, allowing it to form highly stable homotypic or heterotypic interactions (for reviews, see refs 2,11). Based on studies using monoclonal antibodies (mAbs), 1C2 and 1F8, expanded polyglutamine appears to adopt an altered conformation enabling these antibodies to bind preferentially to the pathological glutamine epitope in disease proteins (15,16). The specific nature of this altered conformation remains unclear, but it probably represents a conformational change in the expanded polyglutamine domain or extension of the glutamine domain beyond the normal `shielding' of the rest of the protein, resulting in recognition of the pathological epitope by 1C2 and 1F8.
In this study, we show that non-pathogenic normal ataxin-3 expressed within the nucleus, but not in the cytoplasm, adopts a conformation recognized by the mAb 1C2. Fractionation experiments demonstrate that this novel conformation of ataxin-3 within the nucleus is not the result of proteolysis, suggesting that association with one or more nuclear proteins affects the structure of full-length ataxin-3 which exposes the polyglutamine domain. Further studies confirm that ataxin-3 associates with the nuclear matrix and that 1C2 recognizes ataxin-3 bound to the nuclear matrix. Our data suggest that an early event in the pathogenesis of SCA3/MJD may involve an abnormal conformation of ataxin-3 associated with the nuclear matrix, which exposes the pathological glutamine epitope, allowing it to undergo homotypic and/or heterotypic interactions involved in polyglutamine-mediated neuronal dysfunction.
RESULTS
Expression of ataxin-3 within the nucleus exposes the polyglutamine epitope
Expression constructs used in this study are described in Figure 1. As initially shown by Trottier et al. (15) using western blots, the mAb 1C2 preferentially recognizes the expanded polyglutamine domain of mutant ataxin-3 and other polyglutamine disease proteins. Consistent with this, immunofluorescence analysis of transfected human embryonic kidney (HEK) 293T cells demonstrates that 1C2 recognizes the expanded polyglutamine domain of mutant ataxin-3, whether expressed as part of the full protein or as a C-terminal polyglutamine-containing fragment of ataxin-3 [myc-MJD(78) and NLS-Q78-myc, respectively, in Fig. 2]. Labeling of cells with 1C2, along with an anti-ataxin-3 antibody, shows diffuse nuclear and cytoplasmic labeling for epitope-tagged, expanded repeat ataxin-3, myc-MJD(78) (Fig. 2a). In addition, 1C2 is immunoreactive against the ataxin-3 fragment, NLS-Q78-myc, labeling both the diffuse protein and nuclear inclusions (Fig. 2b).
Figure 1. Expression constructs used in this study. Full-length ataxin-3 containing a normal repeat of 27 glutamines is denoted MJD(27) and expanded repeat ataxin-3 containing 78 glutamines is referred to as MJD(78). In some experiments, normal or expanded repeat ataxin-3 were used containing an N-terminal myc epitope tag [myc-MJD(27) and myc-MJD(78)]. For other experiments, both proteins contained a nuclear localization signal (NLS) at the N-terminus of the protein and a C-terminal myc epitope tag [NLS-MJD(27)-myc and NLS-MJD(78)-myc]. C-terminal fragments of both normal and expanded repeat ataxin-3 were generated such that they include the 14 amino acids proximal to the glutamine repeat, and the remaining C-terminus of the protein including the glutamine domain along with the C-terminal 43 amino acids (Q27 and Q78). The expressed protein contained an N-terminal HA epitope tag (HA-Q27) or an N-terminal NLS and C-terminal myc epitope tag (NLS-Q27-myc or NLS-Q78-myc). Three domains of ataxin-3 constructs are represented: (i) N-terminal 287 amino acids (gray bar); (ii) polyglutamine domain (`Q' bar); and (iii) C-terminal 43 amino acids (stippled bar).
Figure 2. mAb 1C2 binds to pathological forms of ataxin-3. HEK-293T cells were transfected with either full-length expanded ataxin-3, myc-MJD(78) (a), or a pathological fragment of this protein, NLS-Q78-myc (b). In both experiments, cells were double-labeled with anti-ataxin-3 antiserum and 1C2. (a) In cells expressing myc-MJD(78), both 1C2 and anti-ataxin-3 recognize mutant ataxin-3 as a diffuse protein in the nucleus and the cytoplasm. (b) Likewise, both 1C2 and anti-ataxin-3 bind to nuclear and cytoplasmic forms of NLS-Q78-myc in transfected cells. NLS-Q78-myc has a diffuse localization and also forms nuclear inclusions; in both cases, this pathological fragment is recognized by 1C2 as well as by anti-ataxin-3.
Unexpectedly, we found that 1C2 also binds a non-pathological glutamine repeat under two conditions. First, 1C2 binds a truncated fragment of normal ataxin-3 containing 27 glutamines (HA-Q27) when transfected into 293T cells (Fig. 3a). Both 1C2 and an antibody directed at the hemagglutinin (HA) epitope tag show similar diffuse labeling of the HA-Q27 protein throughout the cell (Fig. 3a). This suggests that when isolated from the remainder of the protein, a non-pathological glutamine repeat either acquires a new conformation recognized by 1C2 or becomes accessible to 1C2 because it is no longer `shielded' by the rest of the protein. Second, in some cells expressing full-length ataxin-3 with a repeat of 27 glutamines [myc-MJD(27)], 1C2 selectively stains ataxin-3 within the nucleus (Figs 3b and 4a). Co-labeling with anti-ataxin-3 shows that much of the protein is diffuse and cytoplasmic, yet 1C2 staining is restricted to the nucleus (Fig. 3b). In addition, biochemical fractionation of cells expressing myc-MJD(27) indicates that a majority of the transfected protein is present in the cytoplasmic fraction (Fig. 3c).
Figure 3. mAb 1C2 binds to the non-pathological glutamine domain of truncated ataxin-3 as well as to intranuclear full-length normal ataxin-3. 293T cells were transfected with a C-terminal fragment of normal ataxin-3 (HA-Q27) (a) or the full-length normal protein, myc-MJD(27) (b). (a) Both 1C2 and anti-HA recognize nuclear and cytoplasmic HA-Q27, suggesting that a normal length glutamine domain, isolated from the remainder of the protein, exposes a 1C2 recognizable epitope. (b) In cells expressing myc-MJD(27), anti-ataxin-3 binds to nuclear and cytoplasmic ataxin-3 but 1C2 binds ataxin-3 only in the nucleus in a subset of cells. The nuclear specificity of this immunoreactivity was confirmed by confocal microscopy in which only sections of the cells corresponding to the nucleus (Hoechst-positive) were processed. Note that in these panels, signal through the fluorescein channel (anti-ataxin-3) was decreased to prevent contamination of signal into the rhodamine channel (1C2) making it difficult to detect nuclear ataxin-3. (c) Cellular fractionation of 293T cells expressing myc-MJD(27) confirms that a majority of the transfected protein is expressed within the cytoplasmic fraction. Samples of total cell lysate (T), cytoplasmic (C) and nuclear (N) fractions from transfected cells were probed with anti-ataxin-3.
To study further this selective nuclear labeling of ataxin-3 with the 1C2 antibody, we targeted expression of ataxin-3 to the nucleus through addition of an NLS. Immunofluorescence analysis of cells expressing NLS-MJD(27)-myc shows that increasing expression of ataxin-3 within the nucleus leads to a large increase in the percentage of cells with 1C2 nuclear staining (Fig. 4). The nuclear specificity of 1C2 binding was verified by scanning confocal microscopy of sections through the z-axis of the cell corresponding to the nucleus (Fig. 4a). In cells expressing myc-MJD(27), ~5% of transfected cells demonstrate 1C2 binding in the nucleus, whereas ~45% of cells expressing NLS-MJD(27)-myc have 1C2 binding in the nucleus (Fig. 4b). Approximately 95% of cells positive for 1C2 showed immunoreactivity only in the nucleus, while ~5% showed both cytoplasmic and nuclear immunoreactivity (unpublished data). Similar results were obtained when cells were labeled with mAb 1F8 which also preferentially recognizes pathological polyglutamine domains (16; unpublished data). One possible explanation for the increased nuclear labeling compared with cytoplasmic labeling with 1C2 is that addition of an NLS increases the concentration of ataxin-3 within the nucleus relative to the cytoplasm. However, biochemical fractionation of cells expressing NLS-MJD(27)-myc shows that full-length ataxin-3 is present at roughly equal levels in both the nuclear and cytoplasmic fractions (Fig. 4c; see Discussion). Together, these findings suggest that, within the nucleus of transfected cells, ataxin-3 adopts a conformation similar to HA-Q27 and/or MJD(78), resulting in 1C2 binding.
Figure 4. Targeting expression of full-length normal ataxin-3 to the nucleus increases the percentage of cells displaying nuclear 1C2 immunoreactivity. (a) 293T cells expressing nuclear-targeted ataxin-3, NLS-MJD(27)-myc, were labeled with 1C2, a polyclonal antibody against ataxin-3, NT1 and Hoechst 33342. Only sections corresponding to the nucleus were processed by confocal microscopy. Note that 1C2 binds to normal ataxin-3 when expressed within the nucleus. (b) Cell counts show that in cells expressing myc-MJD(27), ~5% of transfected cells demonstrate nuclear 1C2 binding. However, when nuclear expression is increased through the addition of an NLS, the percentage of transfected cells with nuclear 1C2 immunoreactivity increases to ~45%. (c) Fractionation of cells expressing NLS-MJD(27)-myc demonstrates that the transfected protein is expressed at approximately equal levels in the nucleus and the cytoplasm. Samples of total cell lysate (T), cytoplasmic (C) and nuclear (N) fractions from transfected cells were probed with anti-ataxin-3.
Western blot analysis of 293T cells expressing full-length or truncated forms of ataxin-3 confirms the preferential binding of 1C2 to the expanded glutamine epitope (Fig. 5). Lysates of transfected cells were analyzed by SDS-PAGE followed by western blotting with either an anti-myc antibody directed against the epitope tag (Fig. 5b), or mAb 1C2 (Fig. 5a). Despite loading a greater amount of NLS-MJD(27)-myc (seen on the blot probed with anti-myc), 1C2 binds less efficiently to the normal glutamine repeat compared with the expanded glutamine repeat of NLS-MJD(78)-myc following SDS-PAGE and electrophoretic transfer to PVDF. Identical results were obtained when blots were run in parallel (unpublished data).
Figure 5. Western blot analysis of lysates from transfected 293T cells confirms that 1C2 binds preferentially to expanded polyglutamine domains. Cells were transfected with empty vector, full-length normal or expanded ataxin-3 [NLS-MJD(27)-myc or NLS-MJD(78)-myc], or truncated forms of both proteins (NLS-Q27-myc; NLS-Q78-myc). Samples were run on SDS-PAGE, transferred to PVDF and the blot was first probed with anti-myc (b) then stripped and reprobed with 1C2 (a). A comparison of blots (a) and (b) illustrates that 1C2 preferentially binds the expanded polyglutamine forms of ataxin-3, NLS-MJD(78)-myc and NLS-Q78-myc. Identical results were obtained when blots were run in parallel.
1C2 binding to intranuclear ataxin-3 does not correlate with proteolysis
Limited proteolysis of polyglutamine disease protein is thought to contribute to the pathogenesis of at least some polyglutamine diseases. Many polyglutamine disease proteins, including ataxin-3, have been shown to be substrates for caspases in vitro (17-20). The observation that 1C2 binds ataxin-3 when expressed within the nucleus (but minimally in the cytoplasm) suggests one of two possible scenarios: (i) ataxin-3 is cleaved within the nucleus which liberates a polyglutamine-containing fragment that adopts a conformation similar to that of HA-Q27, thus enabling 1C2 binding; or (ii) ataxin-3 adopts a novel conformation within the nucleus that exposes the polyglutamine epitope, allowing 1C2 to bind the polyglutamine domain within the context of the full-length protein.
To test whether ataxin-3 is cleaved selectively in the nucleus, we performed cellular fractionation of 293T cells expressing myc-MJD(27) or NLS-MJD(27)-myc. Fractions were analyzed by SDS-PAGE followed by western blotting with a series of antibodies against various epitopes of the ataxin-3 protein (Fig. 6). Neither a polyclonal antibody raised against the first 178 amino acids of ataxin-3 (NT1) (21) nor a monoclonal antibody directed against the C-terminal myc epitope tag of NLS-MJD(27)-myc detected a proteolytic fragment in transfected cells (Fig. 6a and b). In addition, we were unable to detect a nuclear proteolytic fragment with polyclonal antiserum raised against full-length ataxin-3 (6), the mAb 1H9 which recognizes amino acids 214-233 of ataxin-3 (22) or 1C2 (Fig. 4c and unpublished data). To confirm the specificity of cellular fractions, blots were probed with anti-histone and anti-tubulin antibodies as markers for the nuclear and cytoplasmic fractions, respectively (Fig. 6c and d). If 1C2 binding is due to proteolytic cleavage of ataxin-3, then protease inhibitors should decrease 1C2 binding. Treatment of cells with leupeptin, p-nitrophenyl-[rho]-guanidinobenzoate, zVAD-fmk or LLNL had no effect on 1C2 binding (unpublished data), consistent with the view that proteolysis is not responsible for 1C2 binding to ataxin-3 in the nucleus.
Figure 6. 1C2 binding to nuclear ataxin-3 is not the result of proteolysis. Cellular fractionation was performed on 293T cells transfected with normal ataxin-3 [myc-MJD(27)] or nuclear-targeted ataxin-3 [NLS-MJD(27)-myc]. In each case, a fixed fraction of total cell lysate (T) was analyzed along with an identical quantity of total protein from the nuclear (N) or cytoplasmic (C) fractions. Cellular fractions were probed with an N-terminal-specific antibody against ataxin-3, NT1 (a), or anti-myc antibody labeling of the epitope tag (b). In either case, ataxin-3 is present at ~50 kDa and there is no detectable proteolytic processing in either the nuclear or cytoplasmic fractions. Samples were also probed with anti-histone antibody (c) or anti-tubulin antibody (d) to verify the purity of cellular fractionation. (c) Histone probe labels the nuclear fractions, while tubulin probe (d) labels the cytoplasmic fractions. Quantity of total protein analyzed for nuclear and cytoplasmic fractions: 5 µg (a and d), 8 µg (b) and 12 µg (c).
It is possible that a very small fraction of ataxin-3 is being cleaved to produce a fragment that is detectable by immunofluorescence but below the level of detection by western blotting. To estimate the limits of our detection, we performed western blots on serial dilutions of full-length normal ataxin-3 as well as on blots of gels overloaded with nuclear protein from cells transfected with NLS-MJD(27)-myc. Using ataxin-3 polyclonal antiserum, we could easily detect 0.6 ng of ataxin-3, whereas the amount of ataxin-3 in overloaded nuclear fractions, where no fragment is observed, was estimated to be 2 µg. Therefore, even if only 0.03% of ataxin-3 is proteolytically cleaved, we should detect it on western blots assuming that the polyclonal antibody recognizes at least one epitope in the fragment as well as the intact protein.
Ataxin-3 with an altered conformation is associated with the nuclear matrix
It has been shown previously that ataxin-3, as well as ataxin-1, associates with the nuclear matrix (10,21). To determine whether the form of ataxin-3 recognized by 1C2 was associated with the nuclear matrix, in situ nuclear matrix preparations were prepared from 293T cells transfected with NLS-MJD(27). These were then processed for immunofluorescence and co-labeled with both anti-ataxin-3 antiserum and 1C2. The form of ataxin-3 recognized by 1C2 was found to be associated with the nuclear matrix (Fig. 7a; ~35% of ataxin-3-positive matrices are 1C2-positive). In addition, experiments employing an alternative method for preparing nuclear matrices (23) confirmed that normal ataxin-3 associates with the nuclear matrix and that the altered conformation of ataxin-3 that binds 1C2 is also associated with the nuclear matrix (unpublished data).
Figure 7. Normal ataxin-3 with an altered conformation and ataxin-3 with an expanded glutamine repeat are associated with the nuclear matrix. In situ nuclear matrix preparations were performed on 293T cells expressing nuclear-targeted normal or expanded repeat ataxin-3, NLS-MJD(27)-myc or NLS-MJD(78)-myc, (a) and (b) respectively. Cells were then double-labeled with anti-ataxin-3 and 1C2. (a) Anti-ataxin-3 labeling confirms that ataxin-3 associates with the nuclear matrix. In addition, 1C2 labeling demonstrates that the altered conformation of nuclear ataxin-3 is retained within the nuclear matrix. (b) Both antibodies label expanded repeat ataxin-3 bound to the nuclear matrix. The nuclear matrix preparation consists of nuclear proteins as well as intermediate filaments (39); ataxin-3 binds to both of these components.
To assess the potential pathological relevance of this association, we examined whether glutamine repeat expansion alters the ability of ataxin-3 to associate with the nuclear matrix. Although it has been shown previously that normal ataxin-3 associates with the nuclear matrix (21), whether mutant ataxin-3 also associates with the nuclear matrix is unknown. To determine this, 293T cells were transfected with NLS-MJD(78)-myc and in situ nuclear matrix preparations were processed for immunofluorescence. Expansion of the glutamine domain in ataxin-3 does not alter its ability to associate with the nuclear matrix (Fig. 7b). The fact that both normal and expanded ataxin-3 associate with the nuclear matrix supports a model of disease in which an early event in pathogenesis may involve an altered conformation of the disease protein in association with the nuclear matrix.
DISCUSSION
The mAb antibody 1C2 preferentially recognizes pathological forms of polyglutamine disease proteins, including the SCA3/MJD disease protein, ataxin-3 (15). 1C2 binding to expanded polyglutamine proteins probably reflects a unique conformation of the expanded glutamine repeat or an extension of the glutamine repeat beyond the `shielding' provided by other domains in the protein. In this study, we demonstrate that in transfected cells 1C2 also recognizes two forms of the normal, non-pathological glutamine domain within ataxin-3: (i) a C-terminal polyglutamine-containing fragment of ataxin-3 with 27 glutamines; and (ii) full-length ataxin-3 with 27 glutamines when it is localized to the nucleus. The unexpected, selective binding of 1C2 to intranuclear ataxin-3 has implications for SCA3/MJD pathogenesis, as it suggests that ataxin-3 undergoes a conformational change in the nucleus that alters the structure and/or accessibility of the polyglutamine domain.
Although it is possible that 1C2 binding to nuclear ataxin-3 reflects a weak binding of 1C2 to full-length ataxin-3 with 27 glutamines combined with a higher concentration of ataxin-3 in the nucleus, data from immunoblots (Fig. 4c) indicate that about the same amount of ataxin-3 is expressed in the nucleus and the cytoplasm. It is still possible that high local concentrations of ataxin-3 present in subnuclear domains are responsible for 1C2 binding. However, high concentrations of purified native ataxin-3 in solution or absorbed to enzyme-linked immunosorbent assay (ELISA) plates are not detected by 1C2, which suggests that a high concentration of ataxin-3 alone, in the absence of the nuclear environment, is not sufficient for 1C2 recognition (unpublished data).
Whereas it does not appear that an increased nuclear concentration of ataxin-3 is responsible for 1C2 binding, it is possible that full-length ataxin-3 is cleaved selectively within the nucleus, forming a C-terminal polyglutamine-containing `Q27-like' fragment, which is then recognized by 1C2. However, using a panel of antibodies directed against various epitopes of the ataxin-3 protein, we were unable to identify proteolytic fragments in either the nucleus or the cytoplasm. In addition, treating cells with various protease inhibitors had no effect on 1C2 binding (unpublished data). These studies argue against a proteolytic event, suggesting instead that a conformational change of the full protein is responsible for exposing the polyglutamine domain and allowing 1C2 binding. This does not, however, rule out a role for proteolysis in the pathology of SCA3/MJD or other polyglutamine diseases. Several polyglutamine disease proteins including ataxin-3 are substrates for caspases (17-19), and large proteins such as huntingtin may require proteolysis for nuclear import (24,25). Proteolysis of disease proteins within the nucleus may be involved in disease progression and/or neural dysfunction (26-28). It will be important to determine whether the apparent conformational change observed with ataxin-3 modulates proteolytic processing or susceptibility.
The fact that the altered conformation of ataxin-3 occurs in the nucleus is intriguing since many models of polyglutamine disease have implicated the nucleus as the primary cellular site for the initiation of neuronal dysfunction. An altered conformation of ataxin-3 within the nucleus that exposes the glutamine domain may represent an early event preceding neuronal dysfunction in SCA3/MJD. One possible sequence of events is that once ataxin-3 is transported to the nucleus, it associates with one or more proteins that either directly alter its conformation or enable it to bind to the nuclear matrix, which then results in an altered conformation that exposes the polyglutamine domain. The exposed polyglutamine domain would then interact with other polyglutamine domains and/or nuclear proteins, forming highly stable complexes that alter both protein and nuclear function, eventually leading to neuronal dysfunction. Alternatively, it is possible that cytoplasmic proteins bind ataxin-3 and block 1C2 binding; during nuclear import, binding to these proteins would be lost. While this is a formal possibility, it seems less likely because purified full-length ataxin-3 with a repeat of 27 glutamines does not bind 1C2 in ELISA assays, which is more consistent with nuclear ataxin-3 adopting a new conformation rather than cytoplasmic ataxin-3 losing a conformation (unpublished data).
Altered protein conformation appears to be the catalyst for initiating cellular dysfunction in other neurodegenerative diseases. The prion disease family is one example in which the normal form of the prion protein, PrPc, is converted into its pathological, protease-resistant form, PrPsc (29). In the case of prion protein, it has been shown that the highly stable PrPsc can convert PrPc from a mostly [alpha]-helical structure to a primarily [beta]-sheet-containing structure. It has been proposed that a cellular protein, factor X, is involved in this conversion possibly by binding to PrPc and helping to alter its conformation (30). The nature of the conversion of ataxin-3 from its normal form to one that binds 1C2 is unknown; however, interactions with specific nuclear proteins would be an obvious possibility. Ataxin-1, the nuclear protein responsible for SCA1, binds leucine-rich acidic nuclear protein (LANP) in a glutamine repeat length-dependent manner, and this interaction may be involved in ataxin-1-mediated pathology (31). Whether ataxin-1 alters its conformation on binding to LANP is unknown; however, they do bind to the nuclear matrix (31). In addition to ataxin-1, huntingtin-interacting proteins, HAP-1, HIP-1 and SH3GL3, have been identified; however, no link has been made between binding to any of these proteins and neuronal dysfunction (32-35).
Although specific nuclear proteins that bind ataxin-3 have not been identified, it is clear that ataxin-3 binds to the nuclear matrix. Ataxin-3 bound to the nuclear matrix appears to have an altered conformation, which suggests one of two possibilities: (i) association of ataxin-3 with the nuclear matrix itself causes a conformational change allowing 1C2 to bind; or (ii) once in the nucleus, ataxin-3 adopts a conformation allowing 1C2 to bind, and subsequently ataxin-3 binds the nuclear matrix. Whichever is the case, binding of ataxin-3 to the nuclear matrix and the apparent conformational change in ataxin-3 are very stable as they are able to withstand detergent and high-salt washes involved in the nuclear matrix preparation.
A key issue that needs to be addressed is whether the apparent conformational change in ataxin-3 that exposes the polyglutamine domain contributes to SCA3/MJD pathogenesis and, more generally, whether it occurs for other polyglutamine disease proteins. Although we have not established a causal link between the conformational change and pathology, the fact that this conformational change occurs selectively in the nucleus and renders the pathogenic protein domain, the glutamine repeat, more exposed or accessible to 1C2 suggests a possible role. The data presented here are based on the normal non-pathological ataxin-3 protein because all forms of pathological ataxin-3 with an expanded polyglutamine domain are recognized by 1C2. It will be important to determine whether ataxin-3 or other polyglutamine disease proteins with an expanded glutamine repeat also adopt a new conformation in the nucleus that exposes the polyglutamine domain. We have shown here that normal ataxin-3 bound to the nuclear matrix has an altered conformation. Ataxin-3 with an expanded glutamine domain also binds to the nuclear matrix, suggesting that pathological ataxin-3 bound to the nuclear matrix may also adopt a conformation that exposes the polyglutamine domain. Exposure of the expanded glutamine repeat of mutant ataxin-3, while associated with the nuclear matrix, may define an early event in the pathogenesis in SCA3/MJD.
MATERIALS AND METHODS
Expression constructs and transfection
The expression constructs used in this study are shown in Figure 1. Myc-MJD(27) and myc-MJD(78) were described previously (36) and represent full-length ataxin-3 with either 27 or 78 glutamines, respectively; each is cloned into the mammalian expression vector pcDNA3. Ataxin-3 with either 27 or 78 glutamines was targeted to the nucleus as previously described (27) through the addition of an NLS and cloned into a modified pAG vector (37) placing a myc-hexahistidine tag at the C-terminus of the protein [NLS-MJD(27)-myc and NLS-MJD(78)-myc, respectively]. An HA-tagged C-terminal fragment of normal ataxin-3 containing 27 glutamines was cloned into the pcDNA3 vector as previously described (9). C-terminal fragments of normal and expanded repeat ataxin-3 (with 27 and 78 glutamines, respectively) containing an NLS were cloned into the modified pAG vector as previously described (36).
Human embryonic kidney (HEK-293T) cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal calf serum and penicillin/streptomycin. One day prior to transfection, cells were plated into 35 mm dishes at a subconfluent density. The next day, cells were transiently transfected as previously described using calcium phosphate (9).
Immunofluorescence and confocal microscopy
At 24 h post-transfection, cells were cultured onto glass coverslips (Fisher Scientific, Pittsburg, PA) pre-treated with 50 µg/ml collagen. Cells were processed for immunofluorescence 12-24 h after plating onto coverslips, by washing in phosphate-buffered saline (PBS) and fixing in 4% paraformaldehyde (pH 6.8) for 10 min. Cells were then rinsed three times in PBS and permeabilized for 10 min in 0.05% Triton X-100 in PBS. Coverslips were next placed in blocking solution (2% goat serum, 0.05% Triton X-100 in PBS) for at least 30 min. Cells were incubated for 90 min at room temperature in the following primary antibodies diluted in blocking solution: 9E10 anti-myc (Calbiochem, San Diego, CA) at 1:100; Y11 anti-HA (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:1000; mAb 1C2 (Chemicon International, Temecula, CA) at 1:15 000; affinity-purified anti-ataxin-3 (6) at 1:1000; or the polyclonal antibody raised against the N-terminus of ataxin-3, NT1 (21), at 1:1000. Cells were then rinsed three times in PBS and incubated for 90 min at room temperature in the following secondary antibodies diluted 1:1000 in PBS: fluorescein isothiocyanate-conjugated goat anti-rabbit (Jackson Laboratories, West Grove, PA) and/or rhodamine-conjugated goat anti-mouse (Jackson Laboratories). In certain experiments, cells were then incubated for 5 min in Hoechst 33342 (Molecular Probes, Eugene, OR) at 2 µg/ml to label chromatin. Lastly, coverslips were rinsed three times in PBS, once in deionized water and mounted on glass slides with Vectashield medium (Vector Laboratories, Burlingame, CA). Confocal images were obtained with a Leica TCS-NT laser confocal imaging system (Heidelberg, Germany) using argon and krypton lasers as well as a UV laser to detect Hoechst-labeled chromatin. Digital images were pseudocolored and processed in Adobe Photoshop.
Cellular fractionation
Cultured cells were fractionated as described previously (15). HEK-293T cells were harvested from 60 mm dishes and rinsed once in PBS. Cell pellets were resuspended in a hypotonic lysis buffer [10 mM HEPES pH 7.9; 1.5 mM MgCl2; 10 mM KCl; 0.5 mM dithiothreitol (DTT)] containing the following protease inhibitors: 2 µg/ml aprotonin, 5 µg/ml leupeptin and 0.5 mM phenylmethylsulfonyl fluoride. Cells were allowed to swell on ice for 20 min and then lysed by passage through a 26-gauge needle. Cell lysis was confirmed by light microscopy and lysis was stopped once ~80-90% of cell membranes were disrupted. Nuclei were then pelleted at 500 gfor 5 min at 4°C. This supernatant was collected and labeled as the cytosolic fraction. The nuclear pellet was then rinsed in lysis buffer and resuspended in a high salt buffer (20 mM HEPES pH 7.9; 420 mM NaCl; 2 mM EDTA pH 8.0; 1.5 mM MgCl2; 25% glycerol; 0.5 mM DTT) containing the protease inhibitors used above. After a 10 min extraction on ice, this fraction was pelleted at 10 000 gat 4°C. This nuclear pellet was resuspended again in the high salt buffer containing 1% Triton X-100 and 0.5% SDS, sonicated briefly, and represents the nuclear fraction. Protein concentrations were determined by Lowry protein analysis (Bio-Rad, Hercules, CA) and an equal amount of protein from each fraction was analyzed by western blot following SDS-PAGE.
Western blotting
HEK-293T cells were rinsed in PBS and resuspended in 100 µl of 2× Laemmli buffer. Lysates were sonicated briefly and heated at 95°C for 3 min prior to electrophoresis on 10% SDS-PAGE. Proteins were then transblotted to PVDF membranes and incubated in blocking solution [5% non-fat milk (Carnation, Glendale, CA) in PBS] for 60 min. Membranes were probed for 60 min with the following primary antibodies diluted in blocking solution: 9E10 anti-myc (Calbiochem) at 1:1000 or mAb 1C2 (Chemicon International) at 1:10 000. For cell fractionation studies, equal amounts of protein were analyzed using the following additional primary antibodies diluted in blocking solution: anti-ataxin-3 antiserum (6) at 1:15 000; NT1 antibody at 1:10 000 (21); anti-tubulin (Sigma, St Louis, MO) at 1:2000; and anti-histone (Boehringer Mannheim, Indianapolis, IN) at 1:160. After incubation in primary antibody, blots were rinsed four times in PBST (0.05% Tween-20 in PBS) for 5 min each and subsequently incubated in block containing either 1:2000 horseradish peroxidase (HRP)-goat anti-rabbit (Santa Cruz Biotechnology) or 1:4000 HRP-goat anti-mouse (Jackson Laboratories) for 60 min. Blots were then rinsed four times in PBST and visualized using enhanced chemiluminescence (Dupont-NEN, Boston, MA).
In situ nuclear matrix preparation
Nuclear matrices were prepared for immunofluorescence using a series of extractions as described previously (38). Transfected 293T cells were split onto collagen-coated glass coverslips 12 h after transfection. The next day, coverslips were rinsed three times in Kern-matrix (KM) buffer (10 mM MES pH 6.2, 10 mM NaCl, 1.5 mM MgCl2, 10% glycerol and 1 µg/ml aprotonin) and then incubated in KM buffer containing 1% NP-40, 1 mM EGTA pH 8.0 and 5 mM DTT for 30 min on ice. Coverslips were rinsed twice in KM buffer and then incubated in KM buffer containing 25 U/ml DNase I (Promega, Madison, WI) for 15 min at 37°C. Next, coverslips were incubated in KM buffer containing 2 M NaCl, 1 mM EGTA and 5 mM DTT for 30 min on ice. Again, coverslips were rinsed twice in KM buffer and then incubated in KM buffer containing 25 U/ml DNase and 50 µg/ml RNase A for 30 min at 37°C. Coverslips were then rinsed in PBS and prepared for immunofluorescence as described previously.
ACKNOWLEDGEMENTS
The authors would like to acknowledge E. Wanker, M. MacDonald and Y. Trottier for their generous contribution of the NT1 antibody against ataxin-3, the 1F8 antibody and the mAb 1H9, respectively. This work was supported by grants from the NIH (R.N.P. and M.K.P.).
REFERENCES
+To whom correspondence should be addressed: Tel: +1 215 898 9736; Fax: +1 215 573 2236; Email: pittman{at}pharm.med.upenn.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  X. Chen, T.-S. Tang, H. Tu, O. Nelson, M. Pook, R. Hammer, N. Nukina, and I. Bezprozvanny Deranged Calcium Signaling and Neurodegeneration in Spinocerebellar Ataxia Type 3 J. Neurosci., November 26, 2008; 28(48): 12713 - 12724. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Alves, E. Regulier, I. Nascimento-Ferreira, R. Hassig, N. Dufour, A. Koeppen, A. L. Carvalho, S. Simoes, M. C. P. de Lima, E. Brouillet, et al. Striatal and nigral pathology in a lentiviral rat model of Machado-Joseph disease Hum. Mol. Genet., July 15, 2008; 17(14): 2071 - 2083. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Zhong and R. N. Pittman Ataxin-3 binds VCP/p97 and regulates retrotranslocation of ERAD substrates Hum. Mol. Genet., August 15, 2006; 15(16): 2409 - 2420. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Purdey The Pathogenesis of Machado Joseph Disease: A High Manganese/Low Magnesium Initiated CAG Expansion Mutation in Susceptible Genotypes? J. Am. Coll. Nutr., December 1, 2004; 23(6): 715S - 729S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Everett and N. W. Wood Trinucleotide repeats and neurodegenerative disease Brain, November 1, 2004; 127(11): 2385 - 2405. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Gonzalez-Alegre and H. L. Paulson Aberrant Cellular Behavior of Mutant TorsinA Implicates Nuclear Envelope Dysfunction in DYT1 Dystonia J. Neurosci., March 17, 2004; 24(11): 2593 - 2601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Chai, S. S. Berke, R. E. Cohen, and H. L. Paulson Poly-ubiquitin Binding by the Polyglutamine Disease Protein Ataxin-3 Links Its Normal Function to Protein Surveillance Pathways J. Biol. Chem., January 30, 2004; 279(5): 3605 - 3611. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Marchal, E. Shehi, M.-C. Harricane, P. Fusi, F. Heitz, P. Tortora, and R. Lange Structural Instability and Fibrillar Aggregation of Non-expanded Human Ataxin-3 Revealed under High Pressure and Temperature J. Biol. Chem., August 22, 2003; 278(34): 31554 - 31563. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Li, T. Macfarlan, R. N. Pittman, and D. Chakravarti Ataxin-3 Is a Histone-binding Protein with Two Independent Transcriptional Corepressor Activities J. Biol. Chem., November 15, 2002; 277(47): 45004 - 45012. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Chai, J. Shao, V. M. Miller, A. Williams, and H. L. Paulson Live-cell imaging reveals divergent intracellular dynamics of polyglutamine disease proteins and supports a sequestration model of pathogenesis PNAS, July 9, 2002; 99(14): 9310 - 9315. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.P.C. van de Warrenburg, R.J. Sinke, C.C. Verschuuren-Bemelmans, H. Scheffer, E.R. Brunt, P.F. Ippel, J.A. Maat-Kievit, D. Dooijes, N.C. Notermans, D. Lindhout, et al. Spinocerebellar ataxias in the Netherlands: Prevalence and age at onset variance analysis Neurology, March 12, 2002; 58(5): 702 - 708. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. O. Evert, I. R. Vogt, C. Kindermann, L. Ozimek, R. A. I. de Vos, E. R. P. Brunt, I. Schmitt, T. Klockgether, and U. Wullner Inflammatory Genes Are Upregulated in Expanded Ataxin-3-Expressing Cell Lines and Spinocerebellar Ataxia Type 3 Brains J. Neurosci., August 1, 2001; 21(15): 5389 - 5396. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Chai, L. Wu, J. D. Griffin, and H. L. Paulson The Role of Protein Composition in Specifying Nuclear Inclusion Formation in Polyglutamine Disease J. Biol. Chem., November 21, 2001; 276(48): 44889 - 44897. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||