Identification and characterization of an ataxin-1-interacting protein ...

© 2000 Oxford University Press

Human Molecular Genetics, 2000, Vol. 9, No. 15 2305–2312

Identification and characterization of an ataxin-1interacting protein: A1Up, a ubiquitin-like nuclear protein Jennifer D. Davidson1,2, Brigit Riley3, Eric N. Burright1,4, Lisa A. Duvick1,4, Huda Y. Zoghbi5 and Harry T. Orr1,2,3,4,+ 1Department

of Genetics, Cell Biology and Development, 2Department of Laboratory Medicine and Pathology, of Biochemistry, Molecular Biology and Biophysics, and 4Institute of Human Genetics, University of Minnesota, Minneapolis, MN 55455, USA, and 5Howard Hughes Medical Institute, Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX 77030, USA

3Department

Received 5 June 2000; Revised and Accepted 7 August 2000

Expansion of a polyglutamine tract within ataxin-1 causes spinocerebellar ataxia type 1 (SCA1). In this study, we used the yeast two-hybrid system to identify an ataxin-1-interacting protein, A1Up. A1Up localized to the nucleus and cytoplasm of transfected COS-1 cells. In the nucleus, A1Up co-localized with mutant ataxin-1, further demonstrating that A1Up interacts with ataxin-1. Expression analyses demonstrated that A1U mRNA is widely expressed as an ∼4.0 kb transcript and is present in Purkinje cells, the primary site of SCA1 cerebellar pathology. Sequence comparisons revealed that A1Up contains an Nterminal ubiquitin-like (UbL) region, placing it within a large family of similar proteins. In addition, A1Up has substantial homology to human Chap1/Dsk2, a protein that binds the ATPase domain of the HSP70like Stch protein. These results suggest that A1Up may link ataxin-1 with the chaperone and ubiquitin– proteasome pathways. In addition, these data support the concept that ataxin-1 may function in the formation and regulation of multimeric protein complexes within the nucleus. INTRODUCTION Spinocerebellar ataxia type 1 (SCA1) is an inherited progressive neurodegenerative disease which primarily affects the brainstem, spinocerebellar tracts and cerebellar Purkinje cells (1). The disease results from the expansion of a polyglutamine stretch within the SCA1 protein, ataxin-1. The glutamines are encoded by an unstable CAG trinucleotide repeat. Studies in transgenic mice expressing expanded human ataxin-1 have demonstrated that the disease is dependent on the nuclear localization of mutant ataxin-1 (2). Furthermore, it has recently been demonstrated that a very early step in SCA1 pathogenesis is the decreased expression of a specific set of genes (3). +To

DDBJ/EMBL/GenBank accession no. AF188240

The functions of both wild-type and mutant ataxin-1 proteins have not been well characterized. One common approach for ascertaining protein function is to identify protein interactors using the yeast two-hybrid system (4,5). Such cDNA library screens have been used to identify interactors of several proteins containing polyglutamine repeat expansions associated with disease. Using this strategy, HIP1, HAP1, cystathionine β-synthase, HYPA, HYPB and HYPC have all been identified as huntingtin-interacting proteins (6–9). The glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase, GAPDH, was shown to associate with ataxin-1, huntingtin, atrophin-1 and the androgen receptor (10,11). The yeast twohybrid system has been used successfully to further characterize ataxin-1. Using this strategy, it was shown that ataxin-1 has the ability to self-associate, regardless of glutamine repeat length, and the domain mediating this interaction was delineated using the yeast two-hybrid system (12). In addition, the nuclear protein LANP was identified as an ataxin-1 interactor utilizing the yeast two-hybrid system (13). LANP is predominantly expressed in nuclei of cerebellar Purkinje cells. Quantitative β-galactosidase assays on co-transformed yeast colonies demonstrated that ataxin-1, harboring an expanded polyglutamine tract, has a stronger interaction with LANP than does wild-type ataxin-1. In this study, we isolated the complete cDNA of a novel ataxin-1-interacting ubiquitin-like protein, designated A1Up. Sequence comparisons revealed that A1Up contains an Nterminal ubiquitin-like region. A1Up also has substantial homology to human Chap1/Dsk2, a protein which binds the ATPase domain of the HSP70-like Stch protein. These results suggest that A1Up may function to link ataxin-1 metabolism with the chaperone and ubiquitin–proteasome pathways. RESULTS Isolation of A1Up To identify proteins that interact with ataxin-1, we performed yeast two-hybrid screens of an adult human brain cDNA

whom correspondence should be addressed. Tel: +1 612 625 3647; Fax: +1 612 626 2600; Email: [email protected]

2306 Human Molecular Genetics, 2000, Vol. 9, No. 15

Figure 1. Localization of the 30-7 encoded fragment of A1U in transfected COS-1 cells. The yeast two-hybrid clone 30-7 was subcloned into pcDNA3.1 (Invitrogen) and expressed in COS-1 cells with and without ataxin-1 constructs. Immunofluorescence was done using either anti-Xpress (Invitrogen) or anti-ataxin-1 antibody 11750. (a) The 30-7 encoded protein exhibited predominantly a diffuse nuclear distribution (red). (b) COS-1 cells co-transfected with 30-7 (red) and ataxin1-30Q (green) demonstrated co-localization of 30-7p and ataxin-1 in nuclear foci. (c) Ataxin-1-82Q (green) inclusions completely sequestered 30-7p (red). (d) Full-length A1Up (red) accumulations co-localized (yellow) with ataxin-1-82Q nuclear accumulations (green).

library (12). These screens revealed 21 distinct cDNAs encoding proteins capable of interacting with ataxin-1. Clone 30-7 was isolated seven times, and examined further based on nuclear localization of the peptide encoded by 30-7 in transfected COS cells (Fig. 1a). In cells co-transfected with 30-7 and ataxin-1-30Q, both were for the most part evenly distributed throughout the nucleus (Fig. 1b). However, in cells cotransfected with 30-7 and ataxin-1-82Q the 30-7 encoded peptide co-localized with the accumulations formed by mutant ataxin-1 (Fig. 1c). By sequence analysis, 30-7 was found to correspond to the 3′ end of an mRNA, encoding only a small portion of the protein. Therefore, a full-length human cDNA clone was isolated, designated A1U (ataxin-1 ubiquitin-likeinteracting protein; GenBank accession no. AF188240) (Fig. 2), and cloned into the mammalian expression vector pcDNA3.1, fusing the coding region to the Xpress epitope tag. Western blot analysis of transfected COS-1 cell extracts using the anti-Xpress antibody revealed a 75 kDa protein (data not shown). Immunofluorescence analysis of transfected COS-1 cells demonstrated a diffuse and punctate distribution within the nucleus and the cytoplasm. In COS-1 cells co-expressing A1U and ataxin-1-82Q or ataxin-1-30Q A1Up co-localized with ataxin-1 (Fig. 1d). No substantial redistribution of ataxin1 (-30Q or -82Q) was seen on co-expression of A1U. Large accumulations of A1Up were often seen adjacent to the ataxin1 inclusions (Fig. 1d). Cells that appeared to be expressing higher levels of ataxin-1-82Q and A1U often had A1Up sequestered within the large ataxin-1 accumulations and/or as a component of the rim of hollow ataxin-1 accumulations. An antibody against A1Up is not currently available, thus it is difficult to assess what alterations in A1Up localization are the result of overexpression.

Interaction of A1Up with ataxin-1 in yeast To characterize more precisely the interaction between A1Up and ataxin-1, β-galactosidase activity was assessed in liquid cultures of yeast (Fig. 3). There was no detectable difference in the production of β-galactosidase between yeast expressing A1Up and ataxin-1-30Q compared with yeast expressing A1Up and ataxin-1-82Q. Thus, A1Up appears to interact equally with wild-type and mutant ataxin-1. However, in yeast expressing A1Up and a deletion variant of ataxin-1-77Q lacking a portion of its self-association region (14), ataxin1∆77Q, the level of β-galactosidase was substantially lower. Thus, the absence of this region from ataxin-1 severely compromises its ability to interact with A1Up in the yeast twohybrid system. Finally, A1Up was found to interact with itself. At least as assessed by the yeast two-hybrid system, the self interaction of A1Up is comparable to that of ataxin-1. A1Up contains regions of homology to other proteins To examine the A1U sequence (GenBank accession no. AF188240) for relationships to other genes and proteins, BLAST comparisons were performed (15). A1U was found to have >99% identity to clone C1orf6 (GenBank accession no. AF113544) located on chromosome 1q21 (16), indicating that A1U is located on chromosome 1q. Additionally, both A1U and C1orf6 contain a 400 bp stretch at their 3′ ends that is identical to the antisense sequence of the 3′ end of the human LIS1 gene on chromosome 17p13 (17). A1U also has significant homology to a cDNA clone of a nuclear protein (GenBank accession no. AB015344) previously described by Ueki et al. (18).

Human Molecular Genetics, 2000, Vol. 9, No. 15 2307

Figure 2. The complete A1U cDNA was assembled using clones 30-7, B10 and K3 as depicted. The portion of A1U sequenced is indicated by the line. The box indicates the A1U coding region. The UbL and UbA domains of A1Up are depicted by the solid boxes within the coding portion. The Sti1-like repeats are depicted by the gray boxes. Locations of the 5′ and 3′ probes used for RNA expression analyses are indicated by the hatched bars.

Chap1/Dsk2 was identified in a yeast two-hybrid screen by virtue of its ability to interact with the ATPase domain of the truncated Hsp-70-like protein Stch (19). The minimal 200 amino acid region of Chap1/Dsk2 required for its binding to Stch contains Sti1-like repeats and a collagen-like proline/ glycine repeat. The A1Up C-terminal region lacks the collagen-like proline/glycine repeat present in Chap1/Dsk2, but it contains two regions that are highly homologous to the Sti1-like repeats of Chap1/Dsk2 (Fig. 4). Amino acids 215– 262 of A1Up are 93% similar (77% identical) to the Sti1-like repeat at amino acids 190–237 of Chap1/Dsk2, and amino acids 458–486 A1Up are 89% similar (78% identical) to the Chap1/Dsk2 Sti1-like repeat at amino acids 433–460. Since a Sti1-like repeat is required for Chap1/Dsk2 binding to Stch (19), the presence of Sti1-like elements in A1Up suggests that it may also bind to the ATPase domain of Hsp70-like proteins. Expression of A1U mRNA in tissues Figure 3. Results from yeast two-hybrid β-galactosidase liquid assays. Reporter gene activity was measured in yeast cultures expressing A1Up and ataxin-1 with either 30 or 82 glutamines, the self interactions of ataxin-1 with 30 or 82 glutamines, the self interaction of A1Up, the interaction of A1Up and ataxin-1 with 77 glutamines lacking a portion of its self-association region (∆77), and the self interaction of ataxin-1-∆77.

Sequence comparisons identified 96 ubiquitin, polyubiquitin, ubiquitin-like and ubiquitin-related proteins with homology to A1Up (data not shown). Each shared homology with a ubiquitin-like (UbL) domain located at the N-terminus of A1Up from amino acid 54 to 88 (Fig. 4). Thus, A1Up is a member of the large family of ubiquitin-like proteins. Significant sequence homology was found throughout the A1Up sequence with the human proteins Chap1/Dsk2 and PLIC-1 (formally ubiquilin-1), and the Xenopus protein XDRP1. Human Chap1/Dsk2 is a duplicated homolog of yeast Dsk2 that is required for proper cell cycle progression (19,20). In addition to the UbL, human Chap1/Dsk2 contains a C-terminal ubiquitin-associated (UbA) domain. At its Cterminus, A1Up contains a 42 amino acid stretch (residues 552–593) that is highly homologous to the UbA of Chap1/ Dsk2.

To assess A1U expression, mRNA expression was examined by dot blot, northern blot and in situ hybridization analyses. We found that A1U was expressed as a single 4.0 kb transcript in murine cerebella (data not shown). A human tissue northern blot was probed with a 5′ probe (Fig. 2) to assess the pattern of A1U mRNA. As in murine cerebella, A1U was expressed as a single 4.0 kb transcript in all tissues examined (Fig. 5a). The highest expression levels were detected in brain, heart, skeletal muscle and pancreas. Weaker expression was observed in kidney, placenta, liver and lung. Transcripts of other sizes were detected using a probe containing the UbL-encoding portion of A1U, presumably due to the UbL homologies (data not shown). The regional pattern of A1U RNA expression within the brain was determined by probing a dot blot of human mRNAs with the 3′ subclone of A1U. The results showed a wide pattern of A1U expression that included the caudate nucleus, putamen, amygdala, whole brain, cerebellum, cerebral cortex, frontal lobe, hippocampus, medulla oblongata, occipital lobe, substantia nigra, temporal lobe, thalamus, nucleus accumbeus, spinal cord and fetal brain (data not shown). Thus, A1U RNA was found to be expressed widely in many tissues and throughout the brain. Given the strong involvement of cerebellar Purkinje cells in SCA1, in situ hybridization was performed to examine A1U


Figure 4. A1Up contains a UbL region and has extensive homology to human Chap1/Dsk2. Dark shading indicates amino acid identities among all proteins and the lighter shading indicates regions of similarity. The UbL, Sti1-like repeats and UbA are indicated by the bars above the sequences. Database comparisons of sequence data were done using NCBI BLAST (15). GenBank accession numbers for each sequence are A1U (AF188240), hPLIC-1 (AF189009), hChap-1/Dsk2 (NM013444) and XDRP-1 (AB030502). Amino acids are depicted using the single letter code.

cell-type expression in murine brain. Sections from a 6-weekold FVB female mouse were probed with the A1U 5′ sense and antisense probes (Fig. 5b–g). The data showed that A1U was expressed in cerebellar granule and Purkinje cells, as well as molecular layer neurons (Fig. 5b). In addition, a high level of

A1U expression was seen in the pyramidal cells of the hippocampus (Fig. 5d), and lower levels in neurons of the frontal cortex (Fig. 5f). No other regions of the brain demonstrated detectable levels of A1U RNA expression by in situ hybridization.


Figure 5. The tissue and cellular expression pattern of A1U RNA. (a) A human multiple tissue northern blot probed with the 5′ subclone of A1U (see Fig. 2). (b–g) The cellular pattern of A1U RNA expression in the brain as assessed by in situ hybridization. Sections from 6-week-old mice hybridized with the A1U 5′ (see Fig. 2), antisense (b, d and f) and sense (c, e and g). High levels of A1U RNA expression were detected in granule and Purkinje cells, and weaker levels of expression in neurons in the molecular layer of the cerebellum (b). Pyramidal cells of the hippocampus exhibited a high level of A1U RNA expression (d). A relatively low level of A1U RNA expression was detected in neurons of the frontal cortex (f).

DISCUSSION Here we report on the isolation and characterization of a novel ataxin-1-interacting protein A1Up. Yeast two-hybrid liquid culture assays indicated that A1Up interacts to a similar extent with wild-type and mutant ataxin-1. Although we have yet to formally localize the region of A1Up responsible for its ability to interact with ataxin-1, the initial yeast two-hybrid identification of the partial A1U cDNA clone, 30-7, encoding a portion of the C-terminus indicates that this region is sufficient for the interaction with ataxin-1. Full-length A1Up was found to localize to punctate structures in the cytoplasm and nucleus of transfected COS-1 cells. Interestingly, the C-terminal portion of A1Up encoded by clone 30-7 localized almost exclusively to the nucleus of transfected cells. This indicates that the sequences of A1Up important for its transport into the nucleus are found within the Cterminal portion. The co-localization of A1Up with ataxin-1 in transfected cells was found to be dependent both on the extent of A1Up expressed and on the number of glutamines in ataxin-1. The most extensive degree of co-localization was found in cells expressing the C-terminal fragment of A1Up and ataxin-1-82Q (Fig. 1c). Co-localization was also seen in cells

expressing full-length A1Up and ataxin-1-82Q (Fig. 1d). In some co-transfected cells, we found that full-length A1Up was sequestered within ataxin-1-82Q accumulations as was seen with the 30-7 encoded fragment of A1Up. Very similar evidence of co-localization was found in cells expressing ataxin-1-30Q and either the C-terminal portion of A1Up or full-length A1Up. Thus, the interaction between A1Up and ataxin-1 may have a role in SCA1 pathogenesis. In this regard, it is interesting to note that the level of interaction between A1Up and the deletion variant of ataxin-1-77Q, ∆77Q, is substantially lower than its interaction with intact ataxin-1 (Fig. 3). Transgenic mice expressing the ataxin-1-∆77Q protein develop Purkinje cell pathology and ataxia very similar to those seen in mice expressing intact ataxin-1-82Q (14). Yet, the disease in aged ataxin-1-∆77 transgenic mice failed to progress to the fullest extent seen in the ataxin-1-82Q mice (P.J. Skinner and H.T. Orr, unpublished data). Perhaps the reduced ability of ataxin-1-∆77Q to interact with A1Up and other proteins alters its ability to induce the full extent of the pathological alterations. The A1U mRNA is 4.0 kb in length and encodes a protein of 601 amino acids. Analysis of the A1Up sequence identified an N-terminal region with substantial homology to the UbL domain found in members of a large family of proteins containing a ubiquitin domain at their N-termini. Among the UbL family of proteins, the function of Rad23 is the best characterized. Rad23 facilitates the formation of a complex between the transcription factor TFIIH and the damage recognition protein RAD14 (21). Rad23 also interacts directly with the 26S proteasome through its N-terminal UbL region (22). Thus, Rad23p provides a link between DNA repair and the ubiquitin–proteasome pathway. Whereas A1Up has not been shown to interact with the 26S proteasome, the presence of an N-terminal UbL suggests that its interaction with ataxin-1 might have a role in the targeting of ataxin-1 to the proteasome. The demonstration of extensive homology with the human Chap1/Dsk2 protein, including the presence of two Sti1-like repeats, has implications for the possible function of A1Up in the nucleus. Sti1-like repeats have been shown to be important for the binding to the ATPase domain of the Hsp70 and Hsp90 (23,24). Moreover, the minimal region of hChap1/Dsk2 required for binding to Stch contains two Sti1-like repeat sequences (19). Another Sti1-like repeat-containing protein, Bag1/Rap46, binds with Hsp70 to inhibit chaperone activity and mediate the formation of multimeric protein complexes including Bcl-2, c-jun, and hormone and growth factor receptors (25–27). The Sti1-like repeats of Chap1/Dsk2 strongly suggest that A1Up may have binding activity for the ATPase domain of chaperones. The Xenopus protein XDRP1 with which A1Up also has substantial similarity, 80% similarity and 78% identity, has been shown to interact with and inhibit the degradation of cyclinA, a mitotic cyclin protein that has also been shown to be involved in nucleation of the mitotic spindle (28). The extensive homology of the ataxin-1-interacting protein A1Up to these proteins, all of which have been demonstrated to participate in the formation and regulation of multimeric protein complexes, some involving the proteasome pathway, suggests that A1Up could have a similar regulatory role involving its interaction with ataxin-1. We have previously demonstrated that in transfected COS cells ataxin-1 associates with the nuclear matrix and that mutant ataxin-1 causes


a redistribution of the nuclear-matrix associated PML protein (29). The demonstration that A1Up, a UbL protein with strong homology to hChap1/Dsk2, hPLIC-1 and XDRP1p, is capable of interacting with ataxin-1 in the nucleus supports the intriguing possibility that through its association with the nuclear matrix and interactions with other nuclear proteins, ataxin-1 has a role in the organization and distribution of multimeric protein complexes in the nucleus. In addition, the ubiquitin–proteasome and chaperone pathways have been shown to have a strong impact on ataxin-1-induced pathology (30,31). Thus, the identification of A1Up as an ataxin-1-interacting protein with a UbL region and Sti1-like repeats, which in other proteins are sites of interaction with Hsp70-like proteins, suggests that A1Up may provide an important role in the regulation of ataxin-1.

2× SSC/0.2% SDS (twice for 15 min each) at room temperature and 0.1× SSC/0.1% SDS (twice for 30 min each at 60°C). Film exposures were at –70°C with intensifying screens. Phage were eluted from picked plugs overnight at 4°C in suspension medium (SM). Secondary and tertiary screens were done in the same manner, except that ∼500 p.f.u. were used per plating. The Rapid Excision kit (Stratagene, La Jolla, CA) was used to isolate the clone from λZAPII. The λgt11 fetal kidney library screens were conducted in a similar manner, except that the host cells were C600-hfi, and inserts were isolated by EcoRI digestion of phage DNAs, followed by subcloning into pBluescript II KS (Stratagene). Isolated clones were sequenced using the Thermo Sequenase Radiolabeled Terminator Cycle Sequencing kit (Amersham Pharmacia, Piscataway, NJ). Assembling the full-length A1U cDNA

MATERIALS AND METHODS Immunofluorescence COS-1 cells were transfected with 5 µg of each construct using the DEAE/dextran method. The following day, cells were divided onto coverslips. After 48 h, cells were fixed in 3.7% formaldehyde, washed three times with phosphate-buffered saline (PBS), permeabilized in acetone at 20°C, rehydrated in PBS, and incubated with primary antibody for 1 h at 37°C. The mouse monoclonal anti-Xpress antibody (1:200; Invitrogen, Carlsbad, CA) was used to detect proteins expressed off the pcDNA3.1 constructs. Ataxin-1 was detected using the 11750 rabbit polyclonal antibody. After three washes in PBS, the cells were incubated with secondary antibody, either an antimouse antibody conjugated with rhodamine or an anti-rabbit antibody conjugated with Cy2 at 1:200 dilution for 1 h at 37°C. The coverslips were mounted onto slides with 4 mg/ml n-propyl gallate in glycerol–gelatin (Sigma, St Louis, MO) after three post-secondary washes in PBS. Analysis was done using confocal microscopy (BioRad, Hercules, CA). cDNA library screening Approximately 2.5 × 104 p.f.u. (per infection) of the pooled B616 λZAPII adult human frontal cortex cDNA library was used to infect XL1-blue MRA cells grown overnight in luria broth, 10 mM magnesium sulfate, 0.2% maltose (37°C for 15 min). Infections were added to NZCYM top agarose (10 g/l NZ amine A, 5 g/l NaCl, 5 g/l bacto-yeast extract, 1 g/l casamino acids, 2 g/l magnesium sulfate, 7 g/l agarose) and poured onto 150 mm NZCYM plates (the same as top agarose, except 15 g/l agarose). After the top agarose had solidified, the plates were inverted and grown overnight at 37°C. The following day, the plates were chilled for at least 1 h prior to doing lifts in duplicate (nylon membrane discs; NEN, Boston, MA). Lifts were air dried for 10 min, followed by denaturation (0.2 M NaOH, 1.5 M NaCl), neutralization (0.4 M Tris–Cl, pH 7.6, 2× SSC), and a 2× SSC wash. Membranes were air dried and baked for 2 h at 80°C. Pre-hybridization was overnight at 42°C in 2× PIPES, 50% formamide, 0.1% SDS and 100 ng/ml denatured salmon sperm DNA. Hybridization was performed the following day using random primer radiolabeled probe in the same solution, temperature and time conditions as prehybridization. Membranes were washed the following day in

The 5′ 722 bp of the A1U coding region was generated by PCR amplifying the sequence from clone K3 (Fig. 2) using the primers 5′A1U (5′-GAATTCGCGGCGGCATGGCGGAGC-3′) and BS-1 (5′-CATCTCTTGCATCATGGC-3′). 5′A1U added an EcoRI site that could ligate in frame into the EcoRI site of pcDNA3.1 (Invitrogen). The PCR product was cloned directly into pCR2.1 (Invitrogen). The 5′ end of A1U was isolated from pCR2.1 by cutting with EcoRI (added to the sequence by PCR) and BspHI (an endogenous A1U restriction site contained within the PCR product). The 3′ BspHI–EcoRI fragment of the B10 clone was isolated from pcDNA3.1/B10. Both fragments were gel purified using Prep-A-Gene (BioRad) and ligated into pcDNA3.1 that had been linearized with EcoRI and treated with alkaline phosphatase. The resulting clones were analyzed for the correct insert size and orientation by digesting with EcoRI and EcoRV, and also double digested with the two enzymes. Clones determined to be correct by restriction analysis were then verified by sequencing the entire region generated by PCR, as well as the 5′ junction. RNA expression analyses A Human RNA Master Blot (Clontech, Palo Alto, CA) was probed with the 3′ A1U probe (Fig. 2). The Human Multiple Tissue Northern (Clontech) was probed with the 5′ A1U (Fig. 2) and the included β-actin control probe as indicated in the protocol provided by the supplier. In situ hybridization analysis was performed on a brain of a 6-week-old female FVB mouse quick frozen in liquid nitrogen and cut into 20 µm coronal sections on a cryostat. The sections were adhered to glass slides, briefly dried on a slide warmer and stored at –70°C until use. Sections were fixed in 4% paraformaldehyde/PBS for 30 min at room temperature and rinsed in PBS. Deproteination was done for 10 min at room temperature (50 mM Tris–Cl pH 8.0, 5 mM EDTA pH 8.0, 0.25 mg/ml proteinase K). The sections were fixed again as described above, followed by a PBS rinse. The sections were acetylated for 10 min at room temperature in 0.25% acetic anhydride, 0.1 M triethanolamine. A PBS rinse was performed prior to dehydrating the sections with a series of ethanol and chloroform incubations (2 min in 70% ethanol, 2 min in 80% ethanol, 2 min in 95% ethanol, 2 min in 100% ethanol, 10 min in chloroform, 5 min in 100% ethanol). Sense and antisense probes were prepared using the DIG RNA labeling kit (Roche, Indianapolis, IN). Hybridization was done at 42°C overnight in 50% formamide, 20 mM Tris–Cl


pH 8.0, 1 mM EDTA, 0.3M NaCl, 1× Denhardt’s solution, 10% dextran sulfate, 500 mg/ml yeast tRNA, after pre-warming the solution and probe at 85°C. The following day, the sections were deparafilmized in 2× SSC for 15 min, washed for 30 min at 42°C in 50% formamide, 2× SSC. Excess probe was removed by incubating at 37°C in RNase buffer (10mM Tris–Cl pH 8.0, 1 mM EDTA, 500 mM NaCl, 200 µg/ml RNase A) for 10 min. Sections were rinsed 10 min each at room temperature in 2× SSC and 1× SSC. A final wash was in 0.5× SSC at 42°C for 30 min. The slides were incubated for 3 h at room temparature in 2× SSC, 0.5% Triton X-100, 4% normal sheep serum. The slides were rinsed three times for 3 min each in buffer A (100 mM Tris–Cl pH 7.5, 150 mM NaCl) at room temparature. The sections were covered with 1:1000 dilution of anti-digoxygenin conjugated to alkaline phosphate (Roche) in 1% normal sheep serum buffer A and set at 4°C overnight. The sections were rinsed at room temparature four times for 10 min each in buffer A followed by a 10 min incubation in buffer B (100 mM Tris–Cl pH 9.0, 100 mM NaCl, 50 mM MgCl2) and covered with coloring solution [buffer B, 2.4 mg/ml levamisole, NBT/ BCIP (Roche)] and incubated overnight at room temperature in a sealed plastic container covered with foil. The reaction was stopped by rinsing the slides in 1× TE pH 8.0 and deionized water, fixed in 4% paraformaldehyde, PBS for 10 min at room temparature and rinsed in deionized water. Following air drying, the slides were dipped in Hemo-De (Fisher, Pittsburgh, PA) and coverslips were mounted over the sections with Permount (Fisher). Yeast two-hybrid analysis Strains and plasmids. Growth and manipulation of yeast strains were done according to standard procedures (32). Yeast plasmids pGBT9, pGAD424, pVA3 (murine p53) and pTD1 (SV40 T antigen) were from Clontech. The plasmid pGBT9 and pGAD424 containing the full-length ataxin-1-30Q, -82Q and -∆77 were constructed as described (12,14). A1U was subcloned into the EcoR1 sites of pGBT9 and pGAD424.

β-galactosidase assays. Yeast with the GAL4 DNA-binding and activation fusion proteins were monitored for β-galactosidase activity using liquid assay methods. Liquid o-nitrophenyl-β-galactoside (ONPG) assays were performed to quantitate two hybrid fusion protein interactions. Y153 cells containing the plasmids of interest were grown overnight in SD medium lacking the appropriate amino acids to maintain selective pressure. Two milliliters of this culture was used to inoculate 10 ml of YPD non-selective medium. Cells were grown at 30°C to OD600 of 0.5–1.0, pelleted, washed in Zbuffer and resuspended in 300 µl Z-buffer. Cells were then permeabilized in liquid nitrogen, thawed, mixed with 700 µl of Z-buffer containing 0.31 µg/ml β-mercaptoethanol and 160 µl of Z-buffer containing 4 mg/ml ONPG. The reactions were incubated at 30°C for between 15 min and 24 h. Reactions were stopped by the addition of 400 µl of 1 M Na2CO3 and the OD420 of the supernatant measured. These assays were performed on a single day, in triplicate, on three independent colonies of each of the co-transformants to control for experimental variation. β-galactosidase units were calculated according to Miller (33).

ACKNOWLEDGEMENTS We thank Dr Marcy MacDonald (MGH) for generously providing the B616 human frontal cortex library, Dr H. Brent Clark (University of Minnesota) for his help in examining in situ hybridization results. This work was supported by grant NS22920 from the NINDS/NIH. REFERENCES 1. Orr, H.T. and Zoghbi, H.Y. (1995) Spinocerebellar ataxia type 1. Semin. Cell Biol., 6, 29–35. 2. Klement, I.K., Skinner, P.J., Kaytor, M.D., Yi, H., Hersch, S.M., Clark, H.B., Zoghbi, H.Y. and Orr, H.T. (1998) Ataxin-1 nuclear localization and aggregation: role in polyglutamine induced disease in SCA1 transgenic mice. Cell, 95, 41–53. 3. Lin, X., Antalffy, B., Kang, D., Orr, H.T. and Zoghbi, H.Y. (2000) Polyglutamine expansion downregulates specific neuronal genes before pathologic changes in SCA1. Nature Neurosci., 3, 157–163. 4. Fields, S. and Song, O. (1989) A novel genetic system to detect protein– protein interactions. Nature, 340, 245–246. 5. Fields, S. and Sterlanz, R. (1994) The two-hybrid system: an assay for protein–protein interactions. Trends Genet., 10, 286–292. 6. Wanker, E.E., Rovira, C., Scherzinger, E., Hasenbank, R., Walter, S., Tait, D., Colicelli, J. and Lehrach, H. (1997) HIP-1: a huntingtin interacting protein isolated by the yeast two-hybrid system. Hum. Mol. Genet., 6, 487–495. 7. Li, X.J., Li, S.H., Sharp, A., Nucifora Jr, F.C., Schilling, G., Lanahan, A., Worley, P., Snyder, S.H. and Ross, C.A. (1995) A huntingtin-associated protein enriched in brain with implications for pathology. Nature, 378, 398–402. 8. Boutell, J.M., Wood, J.D., Harper, P.S. and Jones, A.L. (1998) Huntingtin interacts with cystathionine β-synthase. Hum. Mol. Genet., 7, 371–378. 9. Farber, P.W., Barnes, G.T., Srindidhi, J., Chen, J., Gusella, J.F. and MacDonald, M.E. (1998) Huntingtin interacts with a family of WW domain proteins. Hum. Mol. Genet., 7, 1463–1474. 10. Koshy, B., Matilla, T., Burright, E.N., Merry, D.E., Fischbeck, K.H., Orr, H.T. and Zoghbi, H.Y. (1996) Spinocerebellar ataxia type-1 and spinobulbar muscular dystrophy gene products interact with glyceraldehyde-3phosphate dehydrogenase. Hum. Mol. Genet., 5, 1311–1318. 11. Burke, J.R., Enghild, J.J., Martin, M.E., Jou, Y.S., Meyers, R.M., Roses, A.D., Vance, J.M. and Strittmatter, W.J. (1996) Huntingtin and DRPLA proteins selectively interact with the enzyme GAPDH. Nature Med., 2, 347–350. 12. Burright, E.N., Davidson, J.D., Duvick, L.A., Koshy, B., Zoghbi, H.Y. and Orr, H.T. (1997) Identification of a self-association region within the SCA1 gene product, ataxin-1. Hum. Mol.Genet., 6, 513–518. 13. Matilla, T., Koshy, B., Cummings, C.J., Isobe, T., Orr, H.T. and Zoghbi, H.Y. (1997) The cerebellar leucine-rich acidic nuclear protein interacts with ataxin-1. Nature, 389, 974–978. 14. Klement, I.A., Skinner, P.J., Kaytor, M.D., Yi, H., Hersch, S.M., Clark, H.B., Zoghbi, H.Y. and Orr, H.T. (1998) Ataxin-1 nuclear localization and aggregation: Role in polyglutamine-induced disease in SCA1 transgenic mice. Cell, 95, 41–53. 15. Altschul, S.F., Madden, T.L., Schaffer, A.L., Zhang, J., Zhang, Z., Miller, W. and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., 25, 3389–3402. 16. Fogli, A., Giglio, S., Arrigo, G., Lo Nigro, C., Zollo, M., Viggiano, L., Rocchi, M., Archidiacono, N., Zuffardi, O. and Carrozzo, R. (1999) Identification of two paralogous regions mapping to the short and long arms of human chromosome 2 comprising LIS1 pseudogenes. Cytogenet. Cell Genet., 86, 225–232. 17. Dobyns, W.B., Reiner, O., Carrozzo, R. and Ledbetter, D. (1993) Lissencephaly, a human brain malformation associated with deletion of the LIS1 gene located at chromosome 17p13. J. Am. Med. Assoc., 270, 2838–2842. 18. Ueki, N., Oda, T., Kondo, M., Yano, K., Noguchi, T. and Muramatsu, M. (1998) Selection system for genes encoding nuclear targeted proteins. Nature Biotechnol., 16, 1338–1342. 19. Kaye, F.J., Modi, S., Ivanovska, I., Koonin, E.V., Thress, K., Kubo, A., Kornbluth, and Rose, M.D. (2000) A family of ubiquitin-like proteins binds the ATPase domain of Hsp70-like Stch. FEBS Lett., 467, 348–352.


20. Biggins, S., Ivanovska, I. and Rose, M. (1996) Yeast ubiquitin-like genes are involved in the duplication of the microtubule organizing center. J. Cell Biol., 133, 1331–1346. 21. Guzder, S.N., Bailly, V., Sung, P., Prakash, L. and Prakash, S. (1995) Yeast DNA repair protein RAD23 promotes complex formation between transcription factor TFIIH and DNA damage recognition factor RAD14. J. Biol. Chem., 270, 8385–8388. 22. Schauber, C., Chen, L., Tongaonkar, P., Vega, I., Lambertson, D., Potts, W. and Madura, K. (1998) Rad23 Links DNA Repair to the ubiquitin/proteasome pathway. Nature, 391, 715–718. 23. Höhfeld, J., Minami, Y. and Hartl, F.U. (1995) Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell, 83, 589– 598. 24. Lässle, M., Blatch, G.L., Kundra, V., Takatori, T. and Zetter, B.R. (1997) Stress-inducible, murine protein mSTI1. Characterization of binding domains for heat shock proteins and in vitro phosphorylation by different kinases. J. Biol. Chem., 272, 1876–1884. 25. Takayama, S., Bimston, D.N., Matsuzawa, S., Freeman, B.C., Aime-Sempe, C., Xia, Z., Morimoto, R.L. and Reed, B.C. (1997) BAG-1 modulates the chaperone activity of Hsp70/Hsc70. EMBO J., 16, 4887– 4896. 26. Höhfeld, J. and Jentsch, S. (1997) GrpE-like regulation of the Hsp70 chaperone by the anti-apoptotic protein BAG-1. EMBO J., 16, 6209–6216.

27. Zeiner, M., Gebauer, M. and Gehring, U. (1997) Mammalian protein RAP46: an interaction partner and modulator of 70 kDa heat shock proteins. EMBO J., 16, 5483–5490. 28. Funakoshi, M., Gelay, S., Hunt, T., Takearu, N. and Kobayashi, H. (1999) Identification of XDRP1; a Xenopus protein related to yeast Dsk2p binds to the N-terminus of cyclin A and inhibits its degradation. EMBO J., 18, 5009–5018. 29. Skinner, P.J., Koshy, B.T., Cummings, C.J., Klement, I.A., Helin, K., Servadio, A., Zoghbi, H.Y. and Orr, H.T. (1997) Ataxin-1 with an expanded glutamine tract alters nuclear matrix associated structures. Nature, 389, 971–974. 30. Cummings, C.J., Mancini, M.A., Antalffy, B., DeFranco, D.B., Orr, H.T. and Zoghbi, H.Y. (1998) Chaperone suppression of aggregation and altered subcellular proteasome localization imply protein misfolding in SCA1. Nature Genet., 19, 148–154. 31. Cummings, C.J., Reinstein, E., Sun, Y., Antalffy, B., Jiang, Y.-h., Ciechanover, A., Orr, H.T., Beaudet, A.L. and Zoghbi, H.Y. (1999) Mutation of the E6-AP ubiquitin ligase reduces nuclear inclusion frequency while accelerating polyglutamine-induced pathology in SCA1 transgenic mice. Neuron, 24, 879–892. 32. Sherman, F., Fink, G.R. and Hicks, J.B. (1983) Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 33. Miller, J.H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.