| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Disassembly of nuclear inclusions in the dividing cell-a novel insight into neurodegeneration
Introduction
Results
Nuclear inclusions are highly ordered structures
Ataxin-1 inclusions sequester RED but not vice versa
Inclusion bodies disperse when cells divide
Discussion
Materials And Methods
Acknowledgements
References
Disassembly of nuclear inclusions in the dividing cell-a novel insight into neurodegeneration
Received July 21, 1999; Revised and Accepted September 30, 1999
Spinocerebellar ataxias and Huntington's disease are examples of neurodegenerative diseases caused by a trinucleotide repeat expansion. One hallmark of such diseases is the formation of inclusion bodies (IBs) within neuronal tissue. Although these inclusions may play a pivotal role in the disease process, the reasons underlying their specific accumulation remain obscure. By studying intranuclear IBs in dividing cells we demonstrate for the first time that inclusions such as those of ataxin-1 disperse during mitosis, thus reducing the nuclear aggregate burden. IBs reform in the interphase nucleus. By high-resolution confocal microscopy we also show that inclusions comprise ordered structures capable of homotypic interactions. Unlike those of a non-pathologic protein, ataxin-1 inclusions were shown to be capable of non-specific protein sequestration. Our studies indicate that the specific accumulation of inclusions in terminally differentiated cells such as neurons is a direct consequence of their inability to divide and therefore provides a key to explaining their persistence in neurodegenerative disease.
INTRODUCTION
Abnormal protein aggregation is a common feature of Huntington's disease, spinocerebellar ataxia (SCA) and other glutamine repeat disorders (1). The common genetic defect in such diseases is an expanded polyglutamine tract in a small group of otherwise unrelated proteins (2). Polyglutamine sequences are known to act as activation domains for transcription factors, in vitro (3). Similarly, proteins containing an expanded glutamine tract often show a nuclear localization, and at least one is known to act as a transcription factor (4).
Although the genetic basis of glutamine repeat disorders is now well defined, the role of protein inclusion bodies (IBs) in disease pathology remains controversial (5,6). For example, a protective role might involve sequestration of some toxic soluble component(s). Alternatively, the protein inclusions may themselves be toxic. A pathologic role for IBs would be supported by transgenic mouse studies (7,8) in which animals expressing the neurotoxic fragment of huntingtin were found to develop intranuclear IBs prior to the onset of symptoms. Interestingly, the critical number of repeat units required for disease pathology is identical to that required to drive a conformational change into unusually stable [beta]-sheet structures (2,9,10).
IBs form through the chronic accumulation of misfolded but stable protein. This process can be mimicked by high-level transient expression of proteins within transfected cell lines. Such cellular models recapitulate the cell physiological features of disease and allow the study of IB dynamics using fluorescent fusion proteins. We have recently described a novel human protein called RED that rapidly forms IBs when transiently expressed in cell lines (11) or primary rat neurons (unpublished data). RED contains a repetitive element of 41 alternating acidic/basic amino acids, broken by a single tyrosine (Fig. 1a). These repeats are found in splicing proteins, transcription factors and developmental regulators. Some result from simple reiterations of AGA-GAG codons, possibly as a result of slippage during replication (12). Others, including the repeat found in RED, comprise a varied codon composition implying the need to conserve function. Acidic/basic repeats are associated with genes that cause disease, although the repeat itself has not been shown to cause any pathology. For example, DRPLA encodes a small acidic/basic repeat in addition to a polyglutamine tract that causes a progressive, fatal neurodegenerative disorder (13). The longest acidic/basic repeat that we could identify is in a transcript that maps to the AD3 locus in human chromosome 14. This region is implicated in hereditary Alzheimer's disease, while a physical map of transcripts that co-segregate with AD3 places S164 next to S182, a gene involved in early-onset inherited Alzheimer's disease.
Figure 1. (a) Acidic/basic repeats in RED (EMBL accession no. AJ005579) and its mouse homologue, MURED (EMBL accession no. AJ006130). The RD protein, containing a similar repeat, is part of a transcriptional repressor complex. BLAST searches show that the longest acidic/basic repeat is found in AF109907, a hypothetical protein encoded in the AD3 locus of chromosome 14. This is a region implicated in Alzheimer's disease. U55376 is a Caenorhabditis elegans peptide that contains a reiteration of the AGA-GAG codons to give its extensive RE repeat. (b) The LSCM images of RED IBs. A maximum projection image of RED IBs in the nucleus of a single HT1080 cell is shown. The arrow and arrow-head indicate the base and side of the central IB, respectively. The insert shows an IB containing RED-myc (stained with 9E10 and a cy3 conjugate, which fluoresces red), and the normal achievable resolution of RED-GFP IBs for comparison (lower). Scale bar, 10 µm. (c) Negative image of an amplified section (indicated by a long arrow) of the base of the large IB in (b). The boxed region shows its hexagonal substructure. The edges of these structures are sharpened in (d). Images were manipulated using the Corel Draw package. (e) Optical section through a large IB with the corresponding fluorescence intensity shown below. The peaks in fluorescence indicate the rim; the trough is the hollow.
The acidic/basic repeat of the RED protein and the polyglutamine tracts found in proteins such as ataxin-1 and huntingtin share certain characteristics. Both are polar zippers (14), so called because of their predicted mode of self assembly. They also influence protein folding by favouring the formation of [beta]-sheets (9,14-16). Proteins that contain acidic/basic repeats migrate aberrantly by SDS-PAGE (11,17), while expanded polyglutamines tend to be partially insoluble, leaving an insoluble fraction in the stacking gel during standard SDS-PAGE analysis. Pathologic glutamine zippers accelerate protein aggregation into amyloid fibrils that may sequester other proteins (18-21). One model to explain amyloid formation is that extensive hydrogen bonding between the main-chain and side-chain amides of successive polar zippers form [beta]-sheets. Acidic/basic polar zippers are predicted to drive [beta]-sheet formation in the same fashion (14). Interestingly, a report has shown that insertion of a short polyglutamine sequence into the chymotrypsin inhibitor 2 protein drove it to oligomerize through swapping domains (15). This is a striking demonstration of how a polar zipper may use multiple mechanisms to drive protein aggregation.
This study describes the general behaviour of nuclear IBs using RED and ataxin-1 as examples of non-pathologic and pathologic inclusions. Both proteins were found to form highly ordered inclusions, although only ataxin-1 IBs were capable of non-specific protein sequestration. Importantly, both types of inclusion dispersed when cells entered mitosis, then reformed in the interphase nucleus. This provides a clue as to why IBs can persist in terminally differentiated cells such as neurons.
RESULTS
Nuclear inclusions are highly ordered structures
Initially, we used a green fluorescent protein (GFP) construct of RED as a model system to study the formation and behaviour of nuclear inclusions. Twenty-four hours after transfection, IBs could be discerned as spherical bodies with a brightly fluorescent rim when observed by fluorescence microscopy. These structures appeared grossly similar to those described for cells transfected with ataxin-1, ataxin-3 and huntingtin (18,19,22). Using laser scanning confocal microscopy (LSCM) we were able to observe considerable IB architecture by taking serial optical sections through individual inclusions and visualizing them as maximal projection images (Fig. 1b). IBs consisted of a pseudo-crystalline lattice of pentagonal or hexagonal subunits (Fig. 1c and d). These structures formed in the interphase nuclei of all cell types tested (HT1080, human kidney 293 cells, HeLa, PC12, COS and primary rat neurons) and were seen in both fixed and live cells. Substitution of a myc-tagged construct visualized with the 9E10 antibody ruled out the possibility that these structures might be a GFP artefact (Fig. 1b, insert). Because RED IBs were hollow structures (Fig. 1e) we were able to damp the GFP fluorescence sufficiently to image surface details. RED inclusions were always reticular, and either open (like the large inclusions shown in Fig. 1b) or more spherical. There were no cell-specific differences in inclusion structure. Note that Figure 1b shows possible fusion or blebbing of RED inclusions. (These homotypic interactions were modelled in three dimensions in Fig. 2c). This type of interaction has not been reported previously. RED IBs are not a model for pathologic inclusions. However, the fusion of IBs per se may generate a single large inclusion as often seen in diseased neuronal nuclei and is a dramatic illustration of their interactive potential. Inclusion substructure was also visible under examination by transmission electron microscopy (data not shown). IBs were electron-dense structures with an electron-lucent core, superficially resembling nucleoli, and generally varied from 0.5 to 4 µm in diameter. Occasionally much larger inclusions were formed; these structures are remarkably similar to stigmoid bodies (23) and similar structures of 0.5-3.0 µm in size that have recently been found to contain a huntingtin-associated protein (HAP1) (24,25).
Figure 2. IB of RED[Delta]87-113 in 293 cells. (a) RED for comparison; (b) deletion mutant. The upper images are confocal and the lower images Nomarski. A maximum projection image of RED[Delta]87-113 is shown, with a serial cross section of RED. (c) Surface-filled three-dimensional representation of RED inclusions that demonstrates homotypic interactions (using Zeiss 3D imaging software). This graphics treatment will not show the reticular structure of the inclusion.
Coiled-coil domains are known to drive protein oligomerization (26,27). To investigate the role of the trimeric coiled-coil domain of RED in inclusion formation, we deleted amino acids 87-113 of RED to generate RED[Delta]87-113. This construct was transiently transfected into cells which were examined 24 h later by fluorescence microscopy. RED[Delta]87-113 formed inclusions that contained no visible substructure (Fig. 2b). Optical cross-sections showed that RED[Delta]87-113 inclusions were hollow (data not shown). These bodies formed with the same kinetics as inclusions of the wild-type protein and displayed the same ability as wild-type RED to interact (Fig. 2c).
We subsequently exploited the ability of RED to nucleate IBs as a model system to study non-pathologic inclusions. IBs form after a concentrated burst of high-level expression, with kinetics similar to polyglutamine-containing proteins. The over-expression of RED in these systems, though obviously non-physiologic, is crucial to drive inclusion formation. Likewise, it is standard practice to over-express polyglutamine containing proteins to study inclusion formation. In order to test whether a neurotoxic protein could also self-assemble into IBs with a highly ordered substructure, we then used an SCA-1 construct encoding ataxin-1 with an 82 glutamine repeat. This protein has previously been shown to trigger degeneration of cerebellar Purkinje cells in transgenic mice (28). Ataxin-1 nucleates IBs in transfected cells in a repeat-dependent fashion, i.e. longer repeats nucleate inclusions faster (18). These inclusions are exclusively nuclear and are therefore a suitable pathologic protein with which to compare RED IBs. Cells expressing ataxin-1 show three types of expression pattern: diffuse nuclear inclusions, small punctate nuclear inclusions or few large nuclear inclusions. The propensity to form large inclusions increases with repeat length in a 7:33:60 ratio of diffuse versus punctate versus large inclusions in cells that express ataxin-1 with 82 glutamines (18). High-resolution imaging has not been performed on ataxin-1 inclusions, although these IBs are often as large as nucleoli in transient transfectants (18,21). We found that ataxin-1 transfected into COS-7 and HEK293 cells showed a similar expression profile of inclusion size to that previously reported (18). Figure 3a shows a maximum projection LSCM image of ataxin-1 inclusions. Surface imaging of these inclusions revealed an ordered, honeycomb appearance (Fig. 3a). In contrast to RED IBs, optical sections revealed that the fluorescence intensity of ataxin-1 bodies was equivalent across the structure, suggesting that these inclusions were solid (Fig. 3b). Furthermore, faint fibrillar `feathering' at the periphery of these inclusions was visible by normal fluorescence microscopy, although these cannot readily be appreciated in reproduction. This was not evident in RED inclusions. Clearly, the particular substructures of RED and ataxin-1 inclusions are different. Low-resolution images of ataxin-1 inclusions are no different to those shown previously (18,21,28). It is interesting to note that both RED and ataxin-1 inclusions demonstrate homotypic interactions. Dumb-bell shaped inclusions that may be fusing are shown in Figure 3a. Our experiments suggest that aggregates readily merge. Therefore the seeding of large inclusions by smaller aggregates is a likely scenario. Ataxin-1 IB formation is independent of the GFP tag (18), and is therefore not a construct-specific artefact.
Figure 3. LSCM images of ataxin-1-GFP IBs. (a) Maximum projection image of ataxin-1 inclusions in the nucleus of a single HT1080 cell. (b) Plot of fluorescence intensity through a cross-section of ataxin-1-GFP IBs (shown in top panel).
Ataxin-1 inclusions sequester RED but not vice versa
A key property of pathologic IBs is their capacity to sequester bystander proteins. To study this phenomenon further, we examined heterotypic fusion between non-pathologic RED and pathologic ataxin-1 nuclear inclusions. We were particularly interested in the possibility of interactions between two different polar zipper proteins. Co-expression of a MYC-tagged construct of RED with ataxin-1-GFP generated IBs, visible as green or red inclusions (Fig. 4). Cell nuclei were examined using different filter sets to reveal RED-MYC (Fig. 4b) and ataxin-1-GFP (Fig. 4c). Merged fluorescence is shown in Figure 4a and nuclear staining in Figure 4e. Comparison of Figure 4b and c demonstrates that non-pathologic RED inclusions (Fig. 4b) fail to incorporate ataxin-1. In contrast, each large ataxin-1 inclusion was ringed with RED protein and small inclusions of RED. Unfortunately, because we used an externally added antibody to visualize the myc epitope, the internal composition of the heterotypic bodies could not be resolved. To overcome this limitation we co-expressed ataxin-1-GFP with RED-blue fluorescent protein (BFP). In this case, pure RED IBs and mixed ataxin-1-RED inclusions formed (data not shown). Protein sequestration by ataxin-1 was independent of inclusion size. We also studied the interaction between RED and promyelocytic leukaemia protein (PML). Skinner et al. (18) have shown that PML is sequestered to large inclusions of ataxin-1 that express the expanded glutamine repeat. We found no association between PML and large inclusions of RED (Fig. 4f). Therefore, inclusions nucleated by RED do not sequester either ataxin-1 or PML.
Figure 4. (a) Interaction of RED and ataxin-1. LSCM image of RED-myc and ataxin-1-GFP expressed in COS cells. RED-myc was detected by immunocytochemistry with an anti-myc monoclonal antibody. The second layer is an anti-mouse Cy-3 conjugate that fluoresces red. Separate RED-myc (b) and ataxin-1-GFP (c) images were merged to give (a). The nucleus was stained using DAPI (e) and the exclusion of DNA around ataxin-1 inclusions is shown by the overlain images of (c) and (e) shown in (d). Scale bar, 10 µm. (f) Staining of PML (red) and large RED-GFP inclusions in the nucleus of a transfected HT1080 cell.
Inclusion bodies disperse when cells divide
Despite their wide expression, pathologic CAG repeats exhibit an exquisitely restricted pattern of inclusion formation. The tissues that contain inclusions are often terminally differentiated. We therefore investigated the possibility that inclusions might be destroyed by the process of cell division. Our previous studies (11) indicated that RED inclusions were absent from dividing cells (transient transfection tended to arrest the cell cycle, and we could find very few mitotic cells to examine). Other nuclear bodies are known to redistribute during cell division and inclusions may, therefore, behave in a similar fashion (29). Several strategies were used to view mitotic cells. The first was to express RED-GFP and ataxin-1-GFP in a human cell line (293 embryonic kidney) with relaxed cell cycle checkpoints that readily allow cell cycle progression after transfection. RED and ataxin-1 inclusions were found to disperse throughout the dividing cell (Fig. 5). IBs were occasionally evident at early stages of mitosis but were dispersed by telophase. We could not identify any cells that retained their inclusions through mitosis. Intranuclear aggregates became apparent soon after the nuclear membrane reformed, presumably due to rapid nuclear import. This renucleation of aggregates was captured by time-lapse photography of 293 cells transfected with RED-GFP (Fig. 6).
Figure 5. IBs in dividing cells. Fluorescence photomicrographs of dividing human epithelial kidney cells expressing ataxin-1 and RED. (Top left) Typical prophase cells expressing ataxin-1-GFP using fluorescence. (Middle left) Prophase cell expressing RED-GFP; (bottom left) phase contrast image of the same cell. (Top right) Telophase cells expressing ataxin-1-GFP (inset, ataxin-1-GFP only); (middle and bottom right) telophase cells expressing RED-GFP.
Figure 6. Time-lapse LSCM and phase contrast of RED-GFP in human kidney epithelial 293 cells. Single optical sections were imaged every 5 min. (1) A cell in early mitosis. This division is followed through (8), (9) and (10) until the cells have exited mitosis and are reforming inclusions (13). The last image shows inclusions reformed in one of the cells. The other has detached and is now out of focus.
Another approach was to arrest the cell cycle using nocodazole which inhibits the formation of the spindle apparatus. Nocodazole treatment arrested a large proportion of the cells in early mitosis, the remainder were in interphase. Disordered perinuclear ataxin-1-GFP was abundant in the interphase cells only (Fig. 7). This debris was also seen in non-synchronous cultures, possibly in cells that have already divided. Similar debris in cultures expressing RED-GFP was rare. These findings are consistent with delayed nuclear re-uptake of ataxin-1 and/or degradative processes. Hairpin [beta]-sheet structures, such as those formed by polyglutamines, have been shown to inhibit ubiquitin-dependent protein degradation in vitro. To be degraded by the proteosome a protein must be threaded into the proteolytic chamber. This necessarily involves unfolding, and stable [beta]-hairpin structures may resist these processes (30). Failure to degrade a protein could result in the blockage of the proteosome resulting in a loss of proteolytic activity and an increase in cellular debris.
Figure 7. Ataxin-1-GFP in cytoplasm of nocodazole-treated cells. Inclusions rapidly reform in the nucleus of interphase cells but material trapped in the cytoplasm is disordered (arrows). Diffuse ataxin-1-GFP and small IBs are evident in the cells in early mitosis (m).
DISCUSSION
In this study we present two major findings regarding the behaviour of nuclear protein aggregates. We show, for the first time, that two types of nuclear IB comprise highly ordered structures. The fibrillar inclusions characteristic of polyglutamine aggregates have often been shown in electron micrographs. In this manuscript we demonstrate their three-dimensional aspect and ability to self aggregate. We also demonstrate that nuclear inclusions become dispersed and disassemble when the cell divides. We have deliberately chosen to study and compare inclusions of a pathologic and non-pathologic protein. These may be considered to be models of the physiologic and non-physiologic aggregation described recently by Perutz (31). High-resolution microscopy confirmed that a regular substructure was present in RED and ataxin-1 inclusions. This observation is consistent with the hypothesis that IBs are initially seeded by small aggregates that then drive polymerization into large-scale IBs. Cytoplasmic RED mutants that lack a nuclear localization signal (E. Assier and T. Rich, manuscript in preparation) do not form the ordered inclusions that we see in the interphase nucleus. Therefore, it seems likely that the nuclear environment may be particularly permissive for ordered aggregate formation. By the same reasoning, an alternative polyglutamine-driven folding transition into stable [beta]-sheet structures may also be favoured within a nuclear environment. Nuclear oligomerization of polar zippers into ordered inclusions would represent a classic example of Anfinsen's hypothesis, which states that the tertiary structure of a protein is determined both by its sequence and the aqueous environment in which it finds itself (32).
In dividing cells the nuclear environment is constantly remodelled as the matrix reforms after each cell division. Ataxin-1 and RED both associate with the nuclear matrix, and fluctuations in the matrix composition may induce alternative protein conformations, resulting in the breakdown of inclusions (E. Assier and T. Rich, manuscript in preparation) (18). A trivial explanation for the breakdown of IBs may be that the pathologic protein is diluted when the nuclear membrane dissolves. However, this explanation is unsatisfactory given that polyglutamine-GFP constructs aggregate perfectly well in the cytoplasm of transiently transfected COS-7 cells (33).
The disappearance of IBs following cell division provides an elegant mechanism to explain the toxicity of molecular and macromolecular protein aggregates in non-dividing cells. It has recently been shown that ataxin-1 is especially toxic when expressed in the nucleus (28) and that the nuclear environment promotes aggregation of polyglutamines (19). An N-terminal fragment of huntingtin containing the glutamine repeat has also been shown to seed nuclear inclusions in transfected cells (34,35). We propose that cell division transiently reduces the aggregate burden and, by doing so, delays cell death. Disassembly of sub-microscopic inclusions during cell division may also occur, thus preventing the seeding of further nuclear aggregates and concomitant disease processes. This would be supported by a recent report showing nuclear IBs in post mitotic non-CNS tissue in a mouse model for Huntington's disease (36). The authors concluded that the kinetics of inclusion formation is accelerated in terminally differentiated tissues.
It is unclear whether the mechanisms of cytoplasmic and nuclear IB formation are identical. We have deleted the nuclear localization signals from RED to generate a predominantly cytosolic protein (data not shown). This mutant exhibits a diffuse cytoplasmic expression pattern, although it occasionally forms aggregates close to the nucleus. These small aggregates closely resemble the aggresomes described by Johnston et al. (37). Cytoplasmic aggregation, both of aggresomes and huntingtin, involves the disruption of cytoskeletal elements, in particular intermediate filament protein (37,38), and may be quite different to nuclear aggregation.
In this work we also aimed to explore various aspects of the macromolecular protein organization in inclusions. Coiled-coil domains are known to play a vital role in the protein self-assembly processes. For example, the nuclear mitotic apparatus protein forms a quasi-hexagonal lattice when over-expressed in cell lines (39,40). Shortening of the coiled-coil domain in this protein results in a tightened lattice structure. Similarly, we found that the lattice structure observed for RED inclusions is highly dependent on a trimeric coiled-coil region in the N-terminal portion of the protein. RED[Delta]87-113 inclusions lacking this region exhibited a denser morphology. This may be due to a similar tightening of the normal inclusion structure.
The capacity for organized inclusions to interact with each other, and other proteins, is particularly interesting. Aggregates of ataxin-1 and -3 are known to sequester other proteins, and by doing so have the potential to disturb homeostasis (18,19). We show that large inclusions of RED do not recruit PML or ataxin-1. However, both RED and PML are sequestered by ataxin-1 inclusions. PML is also sequestered by ataxin-3. These data support the notion that pathologic IBs exert their effects through the sequestration of bystander proteins. Sequestration of PML could have the most profound effects. PML has a newly described apoptotic function (41,42), and its disruption may be a common feature of neurodegenerative diseases. PML is expressed in various haematopoietic cell lines and fibroblasts but we could find no reference to PML expression in neurons or neuronal cell lines. Therefore, we stained a neuroblastoma cell line (SK-N-BE-101) for PML protein. Typical promyelocytic oncogenic domains were visualized that localized to the nuclei (Fig. 8).
Figure 8. Fluorescence photomicrograph of PML expression in SK-N-BE101 neuroblastoma cells. PML is stained using the monoclonal antibody PG-M3 and a cy-3 conjugate. The nucleus is stained with DAPI.
Inclusions of pathologic proteins may not be any more toxic than soluble forms, although it is hard to imagine that large inclusions are not detrimental to the cell. However, the macromolecular interactions that we see between IBs and other proteins may reflect molecular processes occurring in solution. The key issue is therefore likely to be the maintenance of a stable nuclear environment that allows aggregates to seed and persist.
MATERIALS AND METHODS
RED was subcloned into pEGFP-N1 (Clontech, Hampshire, UK) as an XhoI-EcoRI insert. Previously we had established that fusion of GFP to either the N- or C-terminus of RED did not alter inclusion formation. The ataxin-1 construct is fused at its N-terminus to GFP. This construct was the gift of Dr H. Zoghbi (HHMI, Baylor College of Medicine, Houston, TX). RED was fused at its C-terminus to a MYC tag in pCDNA3 (Invitrogen, Groningen, The Netherlands). Transient transfection was with Lipofectamine reagent (Gibco BRL, Paisley, UK) according to the manufacturer's instructions. Twenty-four hours post-transfection cells were examined by fluorescence microscopy. Cells were mounted in Vectashield containing DAPI (Vector Laboratories, Peterborough, UK) and examined with a Nikon (Surrey, UK) axiophot microscope using the ×40 or ×100 objective. Immunocytochemistry was with paraformaldehyde-fixed and detergent-permeabilized cells, stained with appropriate primary antibody and Cy-3-conjugated secondary antibody (Sigma, Poole, UK). An anti-PML monoclonal antibody (PG-M3) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Live cells were visualized by rinsing in phosphate-buffered saline then observing directly using the ×40 objective. For three-dimensional reconstruction by LSCM, a series of optical sections through IBs were collected with a Leica (Milton Keynes, UK) TCS-NT laser scanning confocal microscope using an excitation wavelength of 488 nm and a 515/30 band pass filter. The microscope stage was moved incrementally in the z plane through each IB with each increment of movement set at 0.5 × the z resolution of the ×40 objective lens. The images were captured at a resolution of 1024 × 1024 and each frame was averaged four times. The series of images produced was displayed as maximum projections using the Leica `view' software. Time-lapse imaging was with the Zeiss (Welwyn Garden City, UK) LSM-510 confocal microscope using a heated stage. Images were collected every 5 min. Three-dimensional imaging was with standard Zeiss software. Nocodazole treatment was at a concentration of 0.4 µg/ml for 16 h. For examination of mitotic cells, staining with DAPI and [beta]-tubulin was used to assess mitotic status. The nucleotide sequence for RED was previously submitted to the EMBL database under accession no. AJ005579. A murine homologue called MuRED has the EMBL accession no. AJ006130.
ACKNOWLEDGEMENTS
We thank Dr C. Klein for assistance with microscopic imaging techniques. The ataxin-1-GFP construct was a gift from Dr H. Zoghbi. T.R. was supported by the Arthritis Research Council. J.T. was supported by the Wellcome Foundation. The confocal microscopy was carried out in the Multi-Imaging Centre, Cambridge Univeristy, which was established with funding from the Wellcome Trust.
REFERENCES
+To whom correspondence should be addressed. Tel: +44 1223 333921; Fax: +44 1223 333875; Email: tr219{at}mole.bio.cam.ac.uk
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:
|
|
 |
|
  T. Schwarz-Romond, C. Merrifield, B. J. Nichols, and M. Bienz The Wnt signalling effector Dishevelled forms dynamic protein assemblies rather than stable associations with cytoplasmic vesicles J. Cell Sci., November 15, 2005; 118(22): 5269 - 5277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Okuda, H. Hattori, S. Takeuchi, J. Shimizu, H. Ueda, J. J. Palvimo, I. Kanazawa, H. Kawano, M. Nakagawa, and H. Okazawa PQBP-1 transgenic mice show a late-onset motor neuron disease-like phenotype Hum. Mol. Genet., April 1, 2003; 12(7): 711 - 725. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Wyttenbach, J. Swartz, H. Kita, T. Thykjaer, J. Carmichael, J. Bradley, R. Brown, M. Maxwell, A. Schapira, T. F. Orntoft, et al. Polyglutamine expansions cause decreased CRE-mediated transcription and early gene expression changes prior to cell death in an inducible cell model of Huntington's disease Hum. Mol. Genet., August 1, 2001; 10(17): 1829 - 1845. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Sobrido and D. H. Geschwind Molecular Genetics and Inherited Ataxias: Redefining Phenotypes and Pathogenesis Neuroscientist, December 1, 2000; 6(6): 465 - 474. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||