Dentatorubral-pallidoluysian atrophy (DRPLA) is one of the hereditary neurodegenerative disorders caused by expansion of CAG/glutamine repeats. To investigate the normal function of the DRPLA gene and the pathogenic mechanism of neuron death in specific areas of the brain, we isolated and analyzed a gene that shares a notable motif with DRPLA, arginine-glutamic acid (RE) dipeptide repeats. The gene isolated, designated RERE, has an open reading frame of 1566 amino acids, of which the C‐terminal portion has 67% homology to DRPLA, whereas the N-terminal portion is distinctive. RERE also contains arginine-aspartic acid (RD) dipeptide repeats and putative nuclear localization signal sequences, but no polyglutamine tracts. RERE is expressed at a low level in most tissues examined. Immunoprecipitation and invitro binding assays demonstrate that the DRPLA and RERE proteins bind each other, for which one of the RE repeats has a primary role, and extended polyglutamine enhances the binding. With engineered constructs fused with a tag, the RERE protein localized predominantly in the nucleus. Moreover, when RERE is overexpressed, the distribution of endogenous DRPLA protein alters from the diffused to the speckled pattern in the nucleus so as to co-localize with RERE. More RERE protein is recruited into nuclear aggregates of the DRPLA protein with extended polyglutamine than into those of pure polyglutamine. These results reveal a function for the DRPLA protein in the nucleus and the RE repeat in the protein–protein interaction.
Received 4 February 2000; Accepted 22 March 2000.
Dentatorubral-pallidoluysian atrophy (DRPLA) is a progressive neurodegenerative disorder with clinical features of cerebellar ataxia, choreoathetosis, myoclonic epilepsy and dementia. Neuronal loss and gliosis are most severe in the dentate nucleus of the cerebellum and in the globus pallidus, with less profound atrophy in the caudate and putamen (1). We and others (2–4) have detected an unstable expansion of CAG repeats associated with DRPLA. Seven other neurodegenerative disorders including Huntington’s disease, spinobulbar muscular atrophy (SBMA) and spinocerebellar ataxia (SCA 1, 2, 3, 6 and 7) have been shown to be associated with the CAG repeat expansion (5–16). Affected individuals develop symptoms specific to each disorder, which is due to the affected neuronal population in the central nervous system as well as the affected area of the brain.
In each responsible gene, the CAG repeat situates itself in the coding region and encodes polyglutamine. In general, the responsible gene is ubiquitously expressed and the levels of the mRNA and protein expression are not affected by the repeat expansion (17–25). Neuron death in these disorders is believed to result from a pathological gain-of-function arising from the expanded glutamine tract. Overexpression of a small protein carrying extended polyglutamine induces apoptosis in cultured cells where aggregates are formed (26,27). Nuclear inclusion bodies detected in affected brains of patients and model animals seem to be consistent with the aggregates formed in cells by experimental systems (28–32). The molecular mechanism underlying neuron death seems to be common among the CAG/polyglutamine disorders and is attributed to the aggregate formation, or at least to a tendency of aggregation, where extended polyglutamine is a kernel. In contrast, the regional specificity of neurodegeneration has not yet been elucidated. An interaction of the disease gene products with specific partners in selective neuronal populations is known to play a role in determining cell-specific vulnerability. Several interacting proteins for each of the disease gene products have been identified, although their cellular distributions do not strictly correlate with the pathological profiles (33–35). In addition, the normal function of the responsible genes has not been fully elucidated except for the androgen receptor for SBMA and a calcium channel subunit for SCA6.
Characteristic amino acid sequences sometimes provide a clue to function of these genes. We have reported that arginine-glutamic acid dipeptide repeats (RE repeats) occur twice in the C-terminal portion of the DRPLA sequence, and that an expression sequence tag (EST), M78755, has considerable homology to DRPLA in the region spanning the RE repeats (2,3). Although motifs containing RE dipeptide repeats appear in several proteins including 70 kDa snRNP, the function of the repeats has rarely been studied. In contrast, arginine-serine dipeptide repeats (RS repeats) are known to serve as a site for protein–protein interaction, which has been well studied on proteins consisting of spliceosomes (36). Interestingly, the serine residues in the RS repeats are sometimes phosphorylated, resulting in the formation of alternative positive and negative charged residues like the RE repeats (37). Here, we report on a gene represented by EST-M78755. We have demonstrated that the products of this characterized gene and DRPLA interact with each other where the RE repeats have a primary role in the binding. The DRPLA protein is a shuttle plying across the nuclear membrane, and we previously identified a function in a signal transduction pathway coupled with insulin/IGF-1 by detection of IRSp53 as a binding protein with DRPLA (23,38). This report identifies a function for the DRPLA protein in the nucleus in addition to the cytoplasmic function described in the previous report (38).
Characterization of RERE
We determined a cDNA sequence of 8035 bp that was deduced from six overlapping clones and by the 3′ rapid amplification cDNA ends (RACE) method (Fig. 1). It contained an open reading frame of 1566 amino acids (637–5334 nt) and encoded two stretches of RE repeats, thus this gene was designated as RERE (RErepeats encoded). The C-terminal portion of the RERE protein (1104–1566 aa) was very similar to that of the DRPLA protein (732–1185 aa) with 67% identical residues at corresponding positions. In contrast, the N-terminal portion of RERE was distinctive from that of DRPLA. RERE did not have the polyglutamine tract, but had putative nuclear localization signal sequences (NLSs) of bi-partite and mono-partite basic amino acid stretches (39), and also an arginine-aspartic acid (RD) dipeptide repeat. The proximal RE repeat of the DRPLA and RERE proteins was similar in size and almost perfect in the alternative feature of acidic and basic amino acid residues, with some arginine residues being substituted with lysine. The distal RE repeat of DRPLA was relatively short while that in RERE was distorted by insertion of either leucine or isoleucine at every fifth position.
Northern blot analyses detected two transcripts of 7 and 9 kb (Fig. 2). The 7 kb transcript was expressed in all the tissues examined, whereas the 9 kb transcript was expressed exclusively in the pancreas and testis. However, the levels of expression were generally low compared with that of DRPLA, and varied considerably, being moderate in the skeletal muscle, testis and ovary, and very low in the lung, colon and leukocytes.
To detect the RERE product, we first analyzed protein produced in vitro with T7 RNA polymerase and rabbit reticulocyte lysate. A construct carrying the entire coding region produced a protein migrating at an apparent molecular weight of 212 kDa in SDS–PAGE (see below), while the calculation based on the amino acid sequence was predicted to be 172 kDa. This retarded mobility seems due to unusually high proline and serine contents as well as several dipeptide repeats. This phenomenon was also observed in the DRPLA protein (27).
Localization of RERE in the nucleus
As the RERE protein contained putative NLSs, we supposed that RERE was a nuclear protein. To confirm this, we constructed a series of expression constructs in which either full-length or a deleted segment of RERE was fused in-frame with green fluorescent protein (GFP) to generate GFP–full-RERE, C1269, C923, C356, N402, N481 and N1036 (Fig. 3). When GFP–full-RERE was expressed in HeLa cells after transfection, about half of the cells showed a speckled pattern of fluorescence in the nucleus as shown in Figure 3A. Other cells showed a homogeneous distribution outside of the nucleoli in the nucleus [Fig. (3B)], or in both the nucleus and cytoplasm (Fig. 3C). Most products of the deleted construct lacking one of the NLSs (C1269) still localized in the nucleus as full-RERE, while those of other deleted constructs lacking both the NLSs (C923 and C356) primarily formed dots in the cytoplasm, especially in the perinuclear regions, as shown in Figure 3D. Products of the constructs carrying the two NLSs but not the RE repeats (N481 and N1036) localized in the nucleus; however, they did not show the speckled pattern. Thus, the speckled pattern in the nucleus may require the C-terminal portion of RERE as well as the NLS.
Colocalization of the RERE and DRPLA products
We then tested whether RERE colocalized with DRPLA by using the GFP constructs. Although we assumed the DRPLA protein to be a shuttle, it was observed almost uniformly outside of the nucleoli in the nucleus of HeLa cells under the conditions employed (27,38). In transfected HeLa cells expressing GFP–RERE, the distribution of endogenous DRPLA protein altered to the speckled pattern, which almost completely coincided with the presence of GFP–RERE repersented by the yellow region in the merged photos (Fig. 4A–C). This was reflected by a specific interaction between the two molecules since other nuclear-localized endogenous proteins (for example AP-2) did not form the speckled pattern under the same conditions (Fig. 4G–I). The finding was also supported by the fact that the truncated RERE product localized diffusely in the nucleus did not alter the distribution of the DRPLA protein in the nucleus (Fig. 4D–F).
RERE was recruited into aggregates of the DRPLA protein with extended polyglutamine
We then examined the colocalization of the RERE and DRPLA proteins, which were expressed in HeLa cells by cotransfection with two engineered constructs carrying a tag of Xpress or GFP. After staining of transfected cells with an anti-Xpress antibody, a speckled pattern of RERE was detected as observed with GFP–RERE. There were 10–20 dots in each nucleus and their sizes varied from 0.1 to 1 µm. The GFP-tagged full-length DRPLA protein carrying a normal range of polyglutamine localized predominantly in the nucleus and was almost uniformly distributed, as demonstrated previously (27). In cells obtained by cotransfection with Xpress–RERE and GFP–DRPLA-Q14, the distribution pattern of RERE was similar to that observed in transfectants with Xpress–RERE alone (Fig. 5, top). The GFP-tagged, full-length DRPLA protein carrying extended polyglutamine localized in the nucleus like GFP–DRPLA with shorter polyglutamine. However, a fraction (10–20%) of transfected cells produced large aggregates in the nucleus. In such cells, RERE localized in a few relatively larger spots, which coincided with aggregates of DRPLA (Fig. 5, middle). A GFP construct producing a short peptide of extended polyglutamine (GFP–pure-Q71) more easily formed large aggregates in the nucleus. In these cells, the distribution pattern of RERE was not changed significantly, and was similar to that observed in cells transfected only with Xpress–RERE (Fig. 5, bottom). These results, together with the result in Figure 4, indicate that the RERE and DRPLA proteins interact with each other, and that (an) element(s) in addition to polyglutamine is/are involved in the interaction.
Interaction of the RERE and DRPLA proteins detected by immunoprecipitation
Interaction of the RERE and DRPLA proteins was confirmed by immunoprecipitation. We first used cell extracts from 293 cells co-transfected with GFP–DRPLA and Xpress–RERE. Both GFP–DRPLA-Q71 and GFP–DRPLA-Q14 were co-immunoprecipitated with the anti-Xpress antibody, and the amount detected showed a stronger interaction (∼3.3-fold) when polyglutamine was extended (Fig. 6, top). We then tried to detect the interaction of endogenous RERE and DRPLA products, but failed. This was attributed to the low specificity of the anti-RERE antibodies we raised and also to the relatively low abundance of endogenous RERE product in the cells we examined. However, we confirmed the interaction between exogenous RERE and endogenous DRPLA products by immunoprecipitation, coupled with in vivo cross-linking. HeLa cells transfected with GFP–RERE were incubated with a thiol-cleavable chemical cross-linker, dimethyl-3, 3′-dithiobispropionimidate-2HCl (DTBP), and then the cell extracts were subjected to immunoprecipitation with an anti-GFP monoclonal antibody followed by detection with western blotting using an anti-DRPLA antibody. As shown in the bottom panel in Figure 6, endogenous DRPLA protein was detected in precipitates with the antibody for GFP–RERE, but not with a control monoclonal antibody. Truncated forms of RERE, C356 and N1036 did not precipitate endogenous DRPLA protein in similar conditions (data not shown). Although the DRPLA and RERE proteins showed considerable homology, the anti-DRPLA antibody we used was confirmed to have no cross reactivity against endogenous RERE, GFP–RERE or the in vitro translated products of RERE (data not shown).
The interaction was mediated by the RE repeats
Preliminary experiments using the deletion constructs used in Figure 3 indicated that the C-terminal portions of bothRERE and DRPLA were primarily essential to the interaction although polyglutamine also had an influence. To further define the region involved in the interaction, a series of deletion proteins was generated by in vitro translation with pBluescript SK(–) plasmids as indicated in Figure 7, and subjected to an in vitro binding assay. A portion of RERE and DRPLA proteins encompassing the RE repeats was expressed in Escherichia coli as a fused form with glutathione S-transferase (GST), and immobilized on glutathione–Sepharose beads after being purified. Labeled proteins of full-DRPLA, truncated DR-1 and DR-2 bound to both GST–RERE and GST–DRPLA, while DR-3 and DR-4 did not (Fig. 7C and D). Similarly, labeled proteins of full-RERE bound to both GST–DRPLA and GST–RERE, while RE-1, RE-2, RE-3 and RE-4 did not (Fig. 7E and F). All the bound proteins retained at least the proximal RE repeat. To further define the role of the RE repeats in the binding, we made labeled small proteins, carrying either the proximal or distal RE repeat as well as proteins carrying both RE repeats on either the RERE or DRPLA products. These small proteins were then subjected to the binding assay with either GST–DRPLA or GST–RERE. As demonstrated in Figure 7G and H, small proteins carrying at least the proximal RE repeat (D-RE1, D-RE1+2, R-RE1 and R-RE1+2) bound to GST–DRPLA and GST–RERE, whereas small proteins carrying only the distal RE repeats (D-RE2 and R-RE2) did not. Non-labeled synthesized peptides carrying the core sequences of the proximal RE repeat, REREREREKEREREKER or EREREKEKEKEREREREREREAE, abolished the binding, whereas non-labeled peptides carrying the core sequences of the distal RE repeat did not complete the binding (Fig. 7I). These results clearly show that the proximal, not the distal RE, repeat has a major role in the interaction. This implies that RERE and DRPLA proteins form a homodimer in addition to a heterodimer, although we have not proved the homodimer with purified proteins.
We have isolated and characterized the RERE gene, which has sequence homology with the DRPLA gene. The RERE and DRPLA proteins interact with each other, and the proximal RE repeat has a primary role in the interaction. Since we reported the DRPLA sequence with a description of EST-M78755 as an annotation, other groups have determined related nucleotide sequences through cDNA sequencing projects. These reports presented subtle characterization and inconsistencies when compared with our RERE sequence. KIAA0458 (accession no. AB007927) consists of 6642 nt and potentially encodes 1268 aa, where its 180–2280 and 2302–6642 nt sequences correspond to 1465–3565 and 3676–8017 nt (277–974 and 1021–1566 aa) of RERE at a 96–99% level. The sequence connecting the two corresponding portions is different. We have confirmed the integrity of our sequence by RT–PCR and with other ESTs overlapping our sequence (accession nos. AC006991 and AI524637). However, it is still possible that KIAA0458 represents an isoform with a different transcription start site coupled with alternative splicing. The human ARG sequence (accession no. AF016005) contains an open reading frame of 1012 aa in its 4598 nt. Although its 106–4598 nt is almost identical to nt 1839–6330 of RERE, the open reading frame seems to be erroneously set because the sequence lacks the 2140th and 2398th nucleotides of RERE and one nucleotide insertion after 2422 nt. A rat homologue of the RERE gene has been isolated (40). The sequence (accession no. U44091) contains an open reading frame of 1006 aa in its 6659 nt. Its 438–2045 and 2834–6659 nt sequences have homology with 491–2083 and 2084–5952 nt of RERE. The open reading frame of ARP starts at nt 3049 because in‐frame stop codons exist in its 2046–2833 nt region. The extra sequence seems to be intron because the sequences around the junction fit the consensus sequence for splicing. Our study by PCR with human genomic DNA, but not by RT–PCR with mRNA, generated a fragment of ∼800 bp. When removing the extra sequence from U44091, the rat homologue has an open reading frame of 1559 aa which has 92% homology with the human RERE protein.
We have shown that the proximal RE repeat of the RERE and DRPLA proteins serves as the site of interaction in protein–protein binding. As this repeat consists of alternate acidic-basic amino acid residues, the side of α-helix also displays an alternate acidic-basic nature in every seventh residue. Thus, helices carrying the RE repeats may form a coiled-coil structure by the force of opposite electrical charges between two molecules, while polar side chains enforce intertwining in a usual coiled-coil structure. The amino acid sequence around the proximal RE repeat indeed scores high with the Paircoil algorithm (41). As the RE repeats also occur in other proteins including a splicing factor of U1 snRNP, NIP3 (a protein interacting with the nuclear cap binding protein) an uncharacterized RED protein, PQBP-1 (a polyglutamine tract binding protein) and Acinus (an apoptotic factor for chromatin condensation (42–46), the RERE and DRPLA proteins may further interact with these proteins through the RE repeat. In contrast, the distal RE repeats of RERE and DRPLA were not involved in the binding and competition. The distal RE repeat seems not to form a coiled-coil because of its small size in DRPLA and the interruption of alternate positive-negative charges with insertion of leucine/isoleucine in RERE.
Exogenously expressed RERE protein localized mainly in the nucleus, and sometimes formed speckles, which may suggest a specific association with the nuclear matrix ,although we have yet to prove this. Destination of the product to the nucleus is mainly attributed to the distal NLS of the bi-partite basic amino acid stretches as revealed by deletion analysis. In contrast, the speckle formation in the nucleus requires not only the NLS but elements near the C-terminus. The RE repeats are one of the candidate elements but further studies are required to define the others. RERE expression was relatively weak in most tissues and cultured cells as far as we examined by northern blotting and RT–PCR. Since a good antibody is not available at present, the amount and cellular localization of endogenous RERE protein has not been studied. However, the amount of endogenous RERE protein in HeLa cells seems to be very low since the endogenous DRPLA protein was distributed almost uniformly in the nucleus and redistributed into the speckled pattern only when RERE was overexpressed. Localization of the DRPLA protein in the speckles observed in this experiment is due to an unusual imbalance of DRPLA/RERE caused by high-level expression of RERE. Co-transfection with both DRPLA and RERE constructs restored the balance and an almost uniform distribution of the DRPLA protein with normal sized polyglutamine was observed. We have never observed endogenous DRPLA protein to be distributed in a few dots in the nucleus except that carrying extended polyglutamine. Thus, the amount of RERE protein in usual cells is not enough to recruit most of the DRPLA protein, and only a fraction of DRPLA is associated with the RERE protein at a specific site in the nucleus. This may be consistent with the fact that overexpression of the RERE protein is toxic to cells.
Recent studies have revealed that the molecular mechanism underlying neuronal death in CAG/polyglutamine disorders is attributed to the aggregate formation, or at least a tendency for aggregation, where extended polyglutamine plays a central role as a kernel. However, there still remains some controversy over whether nuclear aggregates are vital to, or a consequence of the pathogenesis, although nuclear inclusion bodies are observed both in the patient brain and model animals as well as in cells of experimental systems. Apoptotic cell death of cultured striatum neurons transfected with mutant huntingtin can be prevented by treatment with an inhibitor of caspase, but it failed to suppress the formation of intranuclear inclusions (47). Transgenic mice carrying a deletion within the self-association region of the SCA1 product developed ataxia and Purkinje cell pathology similar to the original SCA1 mice, but nuclear aggregates were not detected in these mice (48). We and others have demonstrated sequential activation of caspase8 and caspase3 by extended polyglutamine (49,50). As these caspases are usually localized in the cytoplasm, the apoptotic process seems to be triggered at a step before polyglutamine enters the nucleus. However, this becomes indistinct because a fraction of these caspases are also localized in the nucleus and processed caspases are detected in the nuclear aggregates (unpublished data). In this regard, this study provides useful information for making a mutant form of the DRPLA protein that lacks the self-association region like the mutant SCA1 product.
It is true that a short peptide containing extended polyglutamine forms aggregates more easily than the whole gene product carrying the same size of polyglutamine tract. However, aggregates detectable under a microscope could be formed even with the whole DRPLA product as shown in this study, which indicates that expression level is another factor involved. In contrast, our in vitro study shows that the interaction of pure extended polyglutamine is not so strong, but the binding becomes tight if there are other interaction sites between a protein and a molecule carrying extended polyglutamine (unpublished data). This also reflects the results shown in Figure 5, where the full-sized DRPLA protein with 71 glutamine repeats formed more condensed aggregates than the pure polyglutamine of 71 residues in the presence of abundant RERE. In this regard, binding proteins to each product of the polyglutamine disorders may reveal the process of not only the site-specific vulnerability of neuron but also aggregate formation.
MATERIALS AND METHODS
Isolation of cDNA clones and DNA manipulation
Plasmid isolation, plaque hybridization, restriction analysis and northern analysis were carried out following standard methods (3,51). We screened two cDNA libraries for the RERE gene; an oligo(dT)-primed human cerebellum cDNA library in a λMaxI vector (Clontech, Palo Alto, CA) and an oligo(dT)- and random-primed human frontal cortex cDNA library in a λZapII vector (Stratagene, La Jolla, CA). Each library was screened by hybridization with a DNA fragment labeled with [32P]dATP by a random primer method. The DNA segment used in the initial screening was derived from the DRPLA gene and contained high homology to EST‐M78755, and segments used in subsequent screening were derived from isolated clones. The rapid amplification of the 3′ ends (RACE) was performed according to the manufacturer’s protocol (Gibco BRL, Rockville, MD). Poly(A) RNA (2 µg) from human adult cerebellum (Clontech, Palo Alto, CA) was reverse-transcribed by a (dT)17 adapter primer and 200 U of SuperScript II RNase H-minus reverse transcriptase (GIBCO BRL). cDNA was then amplified with a primer (5′-GGCCTGTGTCCATTGTAGAT) which corresponds to 6299–6318 nt of RERE. The generated PCR fragment was subcloned onto pUC18. DNA sequencing was carried out by the dideoxy sequencing method using a Sequenase kit, version 2.0 (US Biochemical, Cleveland, OH) with universal primers in pUC18 and pBluescript SK(–)(Stratagene), or with specific internal primers.
Construction of expression plasmids
In order to make a full-length construct of RERE, DNA fragments of SacII–NarI (203–3456 nt), NarI–BfaI (3456–4028 nt) and BfaI–HindIII (4028–5596 nt) from isolated individual cDNA clones were aligned on a plasmid. The AccI–HindIII fragment (580–5596 nt) was then transferred to pBluescript SK(–) to form pBSK-RERE where most of the 5′ non-coding region was removed. The XhoI–HindIII fragment (from a site in the vector to the downstream end) of the pBSK-RERE was transferred into pEGFPC-1 (Clontech) and pcDNA3.1-C (Invitrogen, Groningen, The Netherlands) to generate GFP–RERE and Xpress–RERE, respectively. Truncated GFP–RERE constructs were generated by self-ligation after digestion of GFP–RERE at an appropriate restriction site in the RERE sequence, and at one of the multiple cloning sites in the vector. GST–RERE and GST–DRPLA were constructed by insertion of the SmaI–SmaI fragment (3649–4740 nt) of RERE and the PstI–PstI fragment (2259–3162 nt) of DRPLA onto pGEX-3× (Pharmacia, Uppsala, Sweden). Constructs to produce truncated forms of RERE and DRPLA in the binding assay were generated by subcloning and the end points are described in the legend for Figure 7. Constructs to produce small proteins covering the RE repeats were generated by PCR with either pBSK-DRPLA or pBSK-RERE as a template. The following primers were used which contained an additional EcoRI site in the forward primers and an XbaI site in the reverse primers for convenience in cloning:
5′-GGAATTCCATGGGCTTCAACTCGTGCGCG for D-RE1 and D-RE1+2,
5′-GGAATTCCATGGTGAAGTTGGCTCAGGAG for D-RE2,
5′-GGAATTCCATGGGACCTGGCACCTCGGCCCAG for R-RE1 and R-RE1+2,
5′-GGAATTCCATGGCCATGCCCCACATCAAG for R-RE-2,
5′-GCTCTAGAGCGGGTCCACTGCCCCCAGGGG for D-RE1,
5′-GCTCTAGAGCCGGGCCAGTGGGTCATTGCC for D-RE2 and D-RE1+2,
5′-GCTCTAGAGCATGTGGTAGGCCAGCAGGGG for R-RE1, and
5′-GCTCTAGAGCTCGGCTGCCAGTCTGTCAGG for R-RE2 and R-RE1+2.
Cell culture and transfection
HeLa and 293 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) heat inactivated fetal bovine serum, 50 U/ml penicillin and 0.1 mg/ml streptomycin at 37°C under a humidified atmosphere of 5% CO2. Cells were transfected at 70% confluence in 100-mm dishes or in chamber slides (Nalge Nunc International, Naperville, IL) using Lipofection (Gibco-BRL) according to the protocol recommended by the manufacturer.
Transfected cells on a glass slide were fixed in 4% paraformaldehyde/phosphate buffered saline (PBS) solution (pH 7.4) for 30 min at 4°C, and then permeated with 0.5% Triton X-100/PBS for 30 min at room temperature. Slides were incubated for preblocking with a blotting solution (10 mM Tris, pH 7.6, 150 mM NaCl, 5% skim milk, 2% bovine serum albumin and 0.1% Tween 20) for 1 h at room temperature, and then with an anti-DRPLA (27,38) or anti-AP2 antibody [1:100 diluted with the blotting solution (Santa Cruz Biotechnology, Santa Cruz, CA)] for 1 h at room temperature. After twice washing with PBS for 10 min, bound antibodies were detected with tetramethyl rhodamine isothiocyanate (TRITC)-conjugated swine anti-rabbit IgG [1:20 dilution (Dako, Carpinteria, CA)]. When Xpress-tagged proteins were detected, slides were stained with an anti-Xpress monoclonal antibody [1:100 dilution (Invitrogen)] followed by TRITC-conjugated rabbit anti-mouse IgG [1:20 dilution (Dako)]. Slides were mounted with Vectashield (Vector Laboratories, Burlingame, CA) and viewed by confocal microscopy (Fluoview, Olympus, Tokyo, Japan).
Transfected HeLa cells were harvested with a rubber policeman, washed three times with PBS and resuspended in the lysis buffer (10 mM Tris, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100 and 1% NP-40). Proteins were fractionated by electrophoresis through a 7.5% SDS–polyacrylamide gel and then transferred to a nitrocellulose membrane (Schleicher and Schuell, Dassel). Membranes were incubated in the blotting solution for 2 h at room temperature, and then with a primary antibody at 4°C overnight. Membranes were washed three times with PBS and incubated with a species-specific secondary antibody conjugated with horseradish peroxidase (Sigma, St Louis, MO). Protein bands were visualized with an ECL western blotting detection kit (Amersham, Little Chalfont, UK).
Production of proteins in vitro
Radiolabeled proteins were produced by in vitro transcription–translation reactions with the TNT-coupled reticulocyte lysate system (Promega, Madison, WI) in the presence of [35S]methionine (Amersham), and detected by autoradiography, or quantitatively measured with a phospho-image analyzer (BAS2000; Fuji Film, Tokyo, Japan). GST–DRPLA and GST–RERE fusion proteins were expressed in E.coli (strain BL21) and purified using glutathione–Sepharose 4B (Pharmacia) according to the manufacturer’s protocol.
In vitro protein-binding assays
Purified GST or GST fusion proteins (10 µg) were incubated with the in vitro translated protein products (5–10 µl having the same radioactivity) and 10 µl of glutathione–Sepharose beads in 200 µl of binding buffer (20 mM Tris–HCl, pH 7.5, 100 mM NaCl, 0.2 mM EGTA and 1 mM dithiothreitol) at 4 °C overnight. After washing the beads four times, a bound fraction was eluted by boiling in the SDS–PAGE sample buffer, separated by SDS–PAGE and visualized by autoradiography. For competition assay, the following synthesized peptides (200 µg) were added to the incubation mixture: d-1, REREREREKEREREKER; d-2, AREREREARERDLRDRL; r-1, EREREKEKEKEREREREREREAE; r-2, RERELREREIREREIRERELRER.
In vivo cross-linking
Transfected HeLa cells were washed twice with PBS and then incubated for 20 min at room temperature in PBS with 1 mM DTBP (Pierce, Rockford, IL) with gentle shaking. After extensive washing in PBS, cells were lysed in RIPA buffer (10 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1% deoxycholate and 0.1% SDS) containing protease inhibitors, and immunoprecipitations were performed using an anti-GFP monoclonal antibody (Clontech). Immune complexes were immobilized on protein G beads (Oncogene Science, Uniondale, NY) and washed four times with RIPA, followed by boiling in SDS–PAGE sample buffer. Samples were fractionated by SDS–PAGE followed by immunoblotting using the anti-DRPLA antibody.
We thank A. Asaka and Y. Ohtsuka for technical assistance and K. Saito for manuscript preparation. This study was supported in part by Grants for Genome Research, Brain Research and Pediatric Research from the Ministry of Health and Welfare, a Grant-in-Aid from the Ministry of Education, Science and Culture, and a Grant for Organized Research Combination System from the Science and Technology Agency.
To whom correspondence should be addressed. Tel: +81 3 3414 8121 ext. 2767; Fax: +81 3 3412 2259; Email: firstname.lastname@example.org