Ataxin-3 is subject to autolytic cleavage Pier Luigi Mauri 1 , Matteo Riva 2 , Daniela Ambu 1 , Antonella De Palma 1 , Francesco Secundo 3 , Louise Benazzi 1 , Marco Valtorta 2 , Paolo Tortora 2, * and Paola Fusi 2, * 1 Istituto di Tecnologie Biomediche, CNR, Milano, Italy 2 Dipartimento di Biotecnologie e Bioscienze, Universita ` di Milano-Bicocca, Milano, Italy 3 Istituto di Biocatalisi e Riconoscimento molecolare del CNR, Milano, Italy Spinocerebellar ataxia type 3, also known as Mach- ado–Joseph disease, is one of the nine polyglutamine (polyQ) disorders described to date [1]. They share a number of clinical features, such as late onset and anti- cipation, yet each affects a different area of the brain [1]. Each polyQ disorder is associated with a protein harboring, in its pathologic form, an uninterrupted stretch of polyQ, with a length in excess of a typical threshold [2]. Misfolding of these polyQ-carrying pro- teins has long been known to be associated with the formation of insoluble inclusions [3]. Although such aggregates are a hallmark of neurodegenerative dis- orders, their precise role in the pathologic process remains elusive [4–7]. The polyQ stretch is responsible for protein misfolding, which in turn leads to the gain of a new toxic function, resulting in cell death. This gain of function may account for all the similarities among polyQ diseases [3]. However, it was recently suggested that a loss of function, due to the impair- ment of wild-type protein function caused by the mis- folding and aggregation, could also play a role in the pathogenesis of Huntington disease [8,9]. This led to many studies to define the physiologic and biochemical properties of the nonexpanded proteins. Ataxin-3 (AT-3) is a 42 kDa protein that is respon- sible for spinocerebellar ataxia type 3. In its normal Keywords ataxin-3; mass spectrometry; polyglutamine diseases; proteolysis Correspondence P. Tortora, Dipartimento di Biotecnologie e Bioscienze, Universita ` di Milano-Bicocca, Piazza della Scienza 2, I-20126 Milano, Italy Fax: +39 02 6448 3565 Tel: +39 02 6448 3401 E-mail: paolo.tortora@unimib.it *These authors contributed equally to this work (Received 21 February 2006, revised 15 July 2006, accepted 20 July 2006) doi:10.1111/j.1742-4658.2006.05419.x The protein ataxin-3 is responsible for spinocerebellar ataxia type 3, a neu- rodegenerative disease triggered when the length of a stretch of consecutive glutamines exceeds a critical threshold. Different physiologic roles have been suggested for this protein. More specifically, recent papers have shown that the highly conserved N-terminal Josephin domain of ataxin-3 binds ubiquitin and has ubiquitin hydrolase activity, thanks to a catalytic device specific to cysteine proteases. This article shows that the protein also has autoproteolytic activity, sustained by the same residues responsible for the ubiquitin hydrolase activity. The autolytic activity was abolished when these residues, i.e. Cys14 and His119, were replaced by noncatalytic ones. Furthermore, we found that pretreatment of the protein with tosyl l-phe- nylalanine chloromethyl ketone also abolished this activity, and that this site-specific reagent covalently bound His119, findings supported by MS experiments. MS also allowed us to establish that the attack was aspecific, as cleavage sites were observed at the carboxyl side of apolar, acidic and polar uncharged residues, clustered in the C-terminal, unstructured domain of the protein. In contrast, the Josephin domain was preserved from attack. We propose that the autolytic activity reported here may play a role in pathogenesis, as fragments carrying expanded polyglutamines are thought to be significantly more toxic than the whole protein. Abbreviations AT-3, ataxin-3; AT-3-Q6, murine ataxin-3 carrying six consecutive glutamines; AT-3-Q26, human ataxin-3 carrying 26 consecutive glutamines; EIC, extracted ion chromatogram; polyQ, polyglutamine; GST, glutathione-S-transferase; LC-MS, liquid chromatography coupled to mass spectrometry; LC-MS ⁄ MS, liquid chromatography coupled to tandem mass spectrometry; pCMBS, p-chloro-mercuribenzoate; TLCK, tosyl- L-lysine chloromethyl ketone; TPCK, tosyl-L-phenylalanine chloromethyl ketone. FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS 4277 form, it contains 12–36 glutamines; however, the length of the polyQ in its pathologic variant is in the range 55–84. Its primary structure shows the presence of a conserved N-terminal region, the so-called Jose- phin domain, and of a less conserved C-terminus con- taining the polyQ stretch [10–13]. The increasing interest in the ‘loss of function’ mechanism of pathogenesis has led to many hypothe- ses about the physiologic role of AT-3. Many data suggest that transcriptional repression might be the mechanism for polyQ disorder pathology [14]. On the other hand, a growing number of recent reports sug- gest a role of AT-3 in protein degradation [15,16]. These data are supported by the fact that AT-3 has either two or three potential ubiquitin-interacting motifs, depending on the splice variant [17]. Donald- son et al. [18] also reported that AT-3 is a ubiquitin- binding protein, its functional ubiquitin-binding motifs being required for protein localization into aggregates. In keeping with these observations, through surface plasmon resonance binding analysis, Chai et al. [19] showed that AT-3 is a polyubiquitin-binding protein. Based on a bioinformatics approach, Scheel et al. [20] also recently discovered that the highly conserved Jose- phin domain of AT-3 possesses the predicted catalytic triad of amino acids found in cysteine proteases, puta- tively consisting of Cys14, His119 and Asn134. This led them to predict that AT-3 can remove ubiquitin from polyubiquitinated proteins. Actually, its putative cata- lytic site is similar to that found in ubiquitin C-terminal hydrolase (family 1)-type and ubiquitin C-terminal esterase (ubiquitin thiolesterase)-type ubiquitin proteas- es. This hypothesis was confirmed by replacing the pre- dicted catalytic cysteine (Cys14) by alanine [21], which abolished ubiquitin hydrolase activity towards both the polyubiquitylated 131 I-labeled lysozyme and the fluoro- genic substrate ubiquitin-7-amido-4-methylcoumarin. Although the above-mentioned investigations assigned a ubiquitin hydrolase activity to AT-3, we consistently observed that normal variants of AT-3 underwent slow proteolytic fragmentation when incu- bated at room temperature or even at 4 °C. This led us to check whether AT-3 degradation is autolytic and sustained by the same residues responsible for ubiqu- itin hydrolase activity. For this reason, we replaced the putative catalytic residues by noncatalytic ones, and this abolished the autolytic activity. Although the physiologic significance of this finding has still to be defined, this property may be involved in events lead- ing to pathogenesis, because fragments carrying polyQ stretches are suspected to be more toxic than the whole protein [22]. Results AT-3 carrying six consecutive glutamines (AT-3-Q6) undergoes fragmentation upon incubation at room temperature In previous experiments, we consistently observed that both murine AT-3-Q6 and human AT-3-Q26 under- went slow proteolytic fragmentation when incubated at room temperature for several hours (data not shown). A typical degradation pattern of wild-type murine AT-3-Q6 at 24 °C is shown in Fig. 1 (lanes referred to as wt). HL 9 11DMHL911 D MH L9 1 1WT M DT W CA41T W CA 4 1CA41 h 0h 4 2h 8 4h 4 2 h8 4h0 M 5 4 - 66 - 03 - 4 . 7 9 - 5.12 - 3. 41 - 022 - a b c Fig. 1. SDS ⁄ PAGE (12% gel) of murine ataxin-3 carrying six consecutive glutamines (AT-3-Q6) and its mutants after incubation at 24 °C. After incubation for the indicated times, 6 lg samples of wild-type and mutated proteins were subjected to electrophoresis and revealed by Gel Code staining. DM, double mutant. Lane M shows marker proteins with their relative molecular masses (· 10 )3 ). Bands referred to as ‘a’, ‘b’ and ‘c’ were subjected to in-gel tryptic digestion, and the digestion mixtures analyzed by LC-MS ⁄ MS, which led to their identification as N-terminal fragments covering the regions 1–346, 1–250 and 1–241, respectively. For other details, see Experimental procedures. Autolytic cleavage of ataxin-3 P. L. Mauri et al. 4278 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS C14A, H119L and C14A/H119L mutants of AT-3-Q6 do not undergo proteolytic fragmentation on incubation at room temperature To check whether AT-3-Q6 degradation is autolytic and sustained by the same residues responsible for ubiquitin hydrolase activity, we replaced two putatively catalytic residues, i.e. Cys14 and His119, by noncata- lytic ones, producing the single mutants C14A and H119L and the double mutant C14A ⁄ H119L. The mutants were incubated at 24 °C and subjected to SDS ⁄ PAGE. Figure 1 shows that only faint lower molecular weight bands were detected after incubating C14A for 24 and 48 h, whereas no proteolytic frag- mentation occurred during incubation of both H119L and C14A ⁄ H119L. Far UV-CD spectra of AT-3-Q6 and the mutants collected at 24 °C showed that the two mutations did not significantly affect the overall secondary structure of the protein (Fig. 2). Thus, our results demonstrate a direct involvement of Cys14 and His119 in the autolytic activity. Tosyl- L-phenylalanine chloromethyl ketone (TPCK) inhibits AT-3-Q6 fragmentation Several well-known protease inhibitors, such as phenylmethanesulfonyl fluoride, TPCK, tosyl-l-lysine chloromethyl ketone (TLCK), EDTA, HgCl 2 and p-chloro-mecuribenzoate (pCMBS), were assayed for their ability to prevent the autolytic activity of AT-3- Q6. Except for TPCK, none of them was effective (data not shown). In contrast, TPCK prevented the appearance of proteolytic fragments, as shown by SDS ⁄ PAGE (Fig. 3). To check whether the pattern of covalent modifica- tions of AT-3-Q6 effected by TPCK was consistent with the catalytic role assigned to His119, bands of AT-3-Q6 pretreated with TPCK were excised from the gel and digested with trypsin. Then, the resulting mixtures were analyzed by LC-MS ⁄ MS. This made it possible to identify 81% of the AT-3-Q6 sequence. Moreover, it was shown by LC-MS ⁄ MS analysis that His17, His119 and His198, located in the tryptic peptides T 17)45 ,T 111)124 and T 196)206 , respectively, were modified by TPCK. In contrast, the other histi- dines (His6, His38, His187) were not (Table 1). As a control, the same analysis performed on AT-3-Q6 not subjected to TPCK preincubation did not show any modification of the residue His119 (data not shown). As a representative profile, the MS ⁄ MS spectrum of the TPCK-modified peptide T 111)124 is shown in Fig. 4. Characterization of the large-sized autolytic fragments of AT-3-Q6 To characterize the fragments resulting from autolytic cleavage of AT-3-Q6, the protein was incubated for 48 h at 24 °C and subjected to SDS ⁄ PAGE. Three Fig. 2. Far UV-CD spectra of murine ataxin-3 carrying six consecu- tive glutamines (AT-3-Q6) and the mutants C14A, H119L and C14A ⁄ H119L. Protein samples (0.1 mgÆmL )1 ) were dissolved in 50 m M Tris ⁄ HCl, pH 7.0. The optical pathlength was 0.1 cm. - TPCK M - TPCK + TPCK + TPCK 0 h 48 h 220 - 97.4 - 66 - 45 - 30 - 21.5 - 14.3 - Fig. 3. SDS ⁄ PAGE (12% gel) of murine ataxin-3 carrying six con- secutive glutamines (AT-3-Q6) incubated at 24 °C in the presence or the absence of 1 m M tosyl-L-phenylalanine chloromethyl ketone (TPCK). AT-3 was incubated for the indicated times. Six-microgram samples were then subjected to electrophoresis and revealed by Gel Code staining. Lane M shows marker proteins with relative molecu- lar masses (· 10 )3 ). For other details, see Experimental procedures. P. L. Mauri et al. Autolytic cleavage of ataxin-3 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS 4279 major bands, referred to as ‘a’, ‘b’ and ‘c’, were detec- ted (Fig. 1). They were subsequently subjected to in-gel tryptic digestion, and the digestion mixtures analyzed by LC-MS ⁄ MS. The identified tracts of the sequences represent 68%, 51% and 34.2% of the total, respect- ively. Also, the bands were identified as N-terminal fragments covering the regions 1–346, 1–250 and 1– 241, respectively. Edman degradation of the fragments also confirmed that they carried the expected N-ter- minal sequence resulting from the cleavage of the glutathione-S-transferase (GST)–AT-3 fusion protein (data not shown). On the assumption that cleavage sites autolytically attacked by AT-3-Q6 were different from those attacked by tryptic digestion, we searched for nontryp- tic peptides in the in-gel tryptic digestion mixtures of bands ‘a’, ‘b’ and ‘c’. For this reason, the ‘no-enzyme mode’ of the bioworks software was used, and con- firmed the identification of the C-terminal site D 241 . Surprisingly, we also identified a small number of pep- tides resulting from nontryptic digestion, distant from the putative C-termini of the three fragments under investigation. Non-canonical cleavage sites on protein incubation with trypsin have been already reported, although it is unclear whether this phenomenon is accounted for by a side activity of the protease or by Fig. 4. Fragmentation spectrum (MS ⁄ MS) of the peptide T 111)124 . Peptide identification was achieved by means of SEQUEST data handling. The peptide contains tosyl- L-phe- nylalanine chloromethyl ketone (TPCK)-modi- fied His119. In the inserted table, the calculated masses of B and Y ion fragments are shown, and the matched fragments from the raw spectrum are shown in bold in the table. H* indicates modified His119. Table 1. List of the histidine-containing tryptic fragments from murine ataxin-3 carrying six consecutive glutamines (AT-3-Q6). His- tidines are indicated in bold. Histidines modified by tosyl- L-phenylal- anine chloromethyl ketone (TPCK) are marked with an asterisk. Amino acid numbering starts from the N-terminal methionine of authentic AT-3-Q6. PreScission protease cleavage of the GST–AT-3 fusion protein releases an AT-3 form carrying five residues (GPLGS) upstream of the authentic N-terminus. Peptide fragments Sequence T 1)8 (GPLGS)MESIFHEK T 17)45 H*CLNNLLQGEYFSPVELSSIAHQLDEEER T 35)45 SIAHQLDEEER T 111)124 SFICNYKEH*WFTVR T 196)206 LAH*LKEQSALK T 183)200 VQQMHRPKLIGEELAHLK Autolytic cleavage of ataxin-3 P. L. Mauri et al. 4280 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS additional, contaminating proteolytic activities [23]. Whatever the reason for this observation, it cast no doubt on the identification of the three large-sized fragments described here. Characterization of small-sized autolytic fragments of AT-3-Q6 As shown above, the largest fragments detected in SDS ⁄ PAGE cover the N-terminal region of AT-3-Q6. In order to characterize small-sized polypeptides, which are not retained by the electrophoretic gel, we preincubated preparations of AT-3-Q6 at 24 °C for 48 h and then subjected them to LC-MS analysis with no preliminary tryptic digestion. Polypeptides identi- fied on the basis of their molecular masses cover the C-terminal region only, with sizes in the range 1.7– 9.6 kDa (Table 2). In contrast, no peptide was detected when the mutant H119L was subjected to the same analysis after a 48 h preincubation at 24 °C (data not shown). As a representative profile, Fig. 5 shows the multicharge mass spectrum of peptide 306–355 and the extracted ion chromatograms (EICs) of ion m ⁄ z 784.5 ([M + H + ] 7+ ) in wild-type and C14A mutant sam- ples. Similar results were obtained for the other pep- tides listed in Table 2 (data not shown). These findings also confirm that the observed polypeptides actually result from autolytic activity, and are released by the wild-type protein only, and, on the whole, point to a cleavage pattern involving several sites of autolytic attack. Discussion In recent years, the idea has become widely accepted that many neurodegenerative diseases result not only from the gain of a toxic function of the proteins involved, but also from the loss of their physiologic function [8,9]. This, in turn, led to several investiga- tions aimed at clarifying their physiologic role(s). Different hypotheses have recently been put forward regarding the function of AT-3. Initially, it was sugges- ted that it might play a role in transcriptional regula- tion [14]. However, a growing body of data also points Table 2. Peptides identified by LC-MS of murine ataxin-3 carrying six consecutive glutamines (AT-3-Q6) preparations preincubated for 48 h at 24 °C. For the identification of the peptides, a tool of BIO- WORK software was used. Detected molecular mass Putative AT-3-Q6 region 1731.4 341–355 5484.6 306–355 7414.4 192–257 7565.9 239–302 9553.7 257–341 9555.9 272–355 Fig. 5. Multicharge mass spectrum of the fragmentation peptide T 306)355 obtained by preincubation of wild-type ataxin-3 (AT-3) at 24 °C for 48 h (lower panel). The corresponding extracted ion chromatograms (EICs) of ion m ⁄ z 784.5 ([M + H + ] 7+ ) are shown in the upper panel, along with the C14A mutant sample obtained under the same conditions. P. L. Mauri et al. Autolytic cleavage of ataxin-3 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS 4281 to its possible involvement in proteasome-mediated protein degradation. Actually, AT-3 was shown to bind both ubiquitin, through the two ubiquitin-inter- acting motifs located in its sequence [18], and a num- ber of proteins involved in protein degradation, such as VCP, Rad23 and E4 [24–26]. Furthermore, a recent bioinformatic study suggested that the ubiquitin hy- drolase activity of AT-3 might be sustained by a cata- lytic triad consisting of Cys14, His119 and Asn134 [20]. An ensuing paper provided experimental evidence in support of this hypothesis, in that the mutagenesis of Cys14 abolished ubiquitin hydrolase activity [21]. We consistently observed that pure AT-3 undergoes proteolysis when incubated at room temperature for several hours, and this prompted us to check whether this phenomenon is supported by autolytic cleavage. Based on the hypothesis that the autolytic and ubiqu- itin hydrolase activities are effected by the same resi- dues, we replaced Cys14 and His119 by alanine and leucine, respectively, and this prevented the observed fragmentation. We did not mutagenize residue Asn134, as one cannot expect clearcut results from such replacement. This is because the replacement of the putatively catalytic asparagine by alanine may reduce but does not necessarily abolish enzyme activity, as observed in the case of papain and other cysteine pep- tidases [27]. The occurrence of faint, lower molecular weight bands, even in the electropherograms of the mutated AT-3 forms (Fig. 1), might suggest that the observed fragmentation is at least in part accounted for by con- taminating proteolytic activities. However, this can be definitely ruled out on the basis of the following consid- erations: (a) no time-dependent disappearance of the full-length protein and concomitant accumulation of degradation products was detectable during the incuba- tion of the mutants, as observed in the case of the wild type; (b) although a faint band with size comparable to that of fragment ‘c’ was apparent even in the electro- pherograms of the mutated proteins, additional smaller fragments (in the molecular mass range of about 14– 25 kDa) were consistently detected only in the case of the wild-type protein (Fig. 1); (c) whereas bands ‘a’, ‘b’ and ‘c’ could be identified as fragments of wild-type AT- 3 by in-gel tryptic digestion and LC-MS ⁄ MS analysis, similar analyses run in parallel on the mutated forms did not reveal any such fragment. To further substantiate our conclusions, we used the Escherichia coli protein database (downloaded from the NCBI website) for data handling of mass spectra obtained from wild-type and C14A LC-MS ⁄ MS analy- ses: this allowed us to definitely rule out the presence of any contaminating protease from the microorganism in our samples. Tiny amounts of other contaminants were found instead, including the Hsp70 chaperone protein (data not shown). Our experiments also showed that the autolytic frag- mentation is inhibited by TPCK, and that this site- specific reagent covalently modifies His17, His119 and His198, as supported by MS data. Although we have no obvious explanation for the reactivity of TPCK toward the residues His17 and His198, its ability to covalently bind His119 and at the same time to abolish the self-degradation of the protein is consistent with the above-mentioned hypothesis. The location of the sites of autolytic cleavage (sum- marized in Fig. 6) makes it immediately apparent that no defined specificity can be assigned to this activity, and that these sites are clustered in the C-terminal domain of AT-3, the closest to the N-terminus being next to Leu191. Thus, the Josephin domain, which approximately spans residues 1–182 and is reported as the only one structured in AT-3 [11,12], is completely preserved from proteolytic attack. Our analysis may well have failed to identify all of the cleavage sites: in particular, we observed slightly different cleavage pat- terns in different experiments. However, this does not contradict the above-mentioned conclusions, as proteo- lytic fragmentation was never observed outside the C-terminal, unstructured domain. As the identified Fig. 6. Location of the autolytic cleavage sites of murine ataxin-3 carrying six con- secutive glutamines (AT-3-Q6). Arrows indi- cate the identified cleavage sites. The Josephin domain is shaded. Autolytic cleavage of ataxin-3 P. L. Mauri et al. 4282 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS sites of cleavage are at the carboxyl side of apolar, aci- dic and polar uncharged residues, it seems unlikely that cleavage also occurs next to basic residues, in keeping with the lack of any inhibitory effect by TLCK. Furthermore, as only one site close to an aci- dic residue was identified (i.e. Asp241), we conclude that AT-3 activity is not caspase-like. The ability of AT-3 to undergo autolytic degradation might be related to the mechanisms of pathogenesis, as proteolysis is widely thought to play a role in the cas- cade of events that eventually lead to the onset of many neurodegenerative diseases [28]. In particular, fragments containing polyQ stretches have been reported to be more toxic than the whole proteins [29–34], although their proteolytic fragmentation is generally ascribed to caspases. In vivo studies also showed that expanded AT-3 undergoes proteolysis, with the resulting formation of a cytotoxic, polyQ-containing fragment [35]. Possibly, autolysis might play an additional role in releasing polyQ-containing, pathogenic fragments in vivo. At this stage, it cannot be established whether the observed fragmentation is involved in physiologic pro- cesses such as protein turnover or nuclear localization. It should be pointed out, however, that a recent paper presenting the Josephin architecture of AT-3 solved by NMR spectroscopy showed that it bears a clear simi- larity to members of the papain-like cysteine proteases [36]. This finding nicely confirms our results, and fur- ther supports the idea that the proteolytic activity that we assigned to AT-3 may play a still unidentified physiologic role. Finally, the fact that AT-3 is subject to autolysis also has methodologic relevance. Actually, this prop- erty, along with the occurrence of the unstructured C-terminal domain, explains why attempts to crystal- lize the protein have been unsuccessful so far. We expect that this goal may be achieved when the isola- ted Josephin domain is subjected to crystallization tri- als, where Cys14 and ⁄ or His119 have been replaced by noncatalytic residues. Experimental procedures Gene cloning and mutagenesis Murine AT-3-Q6 was cloned into plasmid pGEX6P-1, from its cDNA identified in the EST database (dbEST), as previ- ously reported [37]. The recombinant protein was then expressed as a GST fusion protein in the E. coli Codon Plus-RIL strain, and retrieved by means of a PreScission Protease (Amersham Biosciences UK Ltd, Little Chalfont, UK) cleavage site located in-between the two coding sequences. Murine AT-3-Q6 mutagenesis to produce C14A, H119L and C14AH119L was carried out with the Quick Change Mutagenesis Kit (Stratagene, La Jolla, CA). According to the manufacturer’s instructions, specific prim- ers carrying the mutated codon were used to produce and amplify the mutated plasmid by means of PCR. Parental DNA was then digested with DpnI endonuclease for 1 h at 37 °C, and E. coli XL-Blue cells were transfected with recombinant plasmids (1 lL of the reaction mixture was used to transfect 50 lL of supercompetent XL-Blue cells). Protein purification All proteins were expressed in E. coli strain BL21 Codon Plus RIL as GST fusion proteins containing a PreScission Protease recognition site. Cells were grown at 37 °Cin LB ⁄ ampicillin medium and induced with 50 lm isopropyl thio-b-d-galactoside (IPTG) at A 600 ¼ 1.0 for 2 h. To obtain crude extracts, cells were frozen and thawed, incuba- ted for 1 h at 4 ° C in 100 mL of lysis buffer (10 mm sodium phosphate, pH 7.2, 150 mm NaCl, 1 mm phenyl- methanesulfonyl fluoride, 10 mm dithiothreitol, 100 mm MgCl 2 , 0.5 mg mL )1 lysozyme), and again frozen and thawed. Triton X-100 (1%) and DNase [0.2 mgÆ(g cells) )1 ] were then added, and the samples were further incubated for 30 min at room temperature. Finally, they were centri- fuged for 30 min at 47 000 g in an Avanti J-20 centrifuge with a JA20 rotor (Beckman Coulter, Fullerton, CA). The supernatants were incubated in batches with Glutathione Sepharose 4B affinity resin [0.1 mLÆ(mL crude extract) )1 ] for 45 min at room temperature. The resin was subse- quently packed into a column and washed with three vol- umes of 10 mm sodium phosphate (pH 7.2) and 150 mm NaCl, and then three volumes of cleavage buffer (50 mm Tris ⁄ HCl, pH 7.0, 150 mm NaCl, 1mm EDTA, 1 mm di- thiothreitol). Removal of the GST affinity tail was achieved by incubating the resin-bound proteins overnight at 4 °Cin the presence of PreScission Protease [400 UÆ(mL resin) )1 ]. Mature AT-3-Q6 proteins were then eluted with cleavage buffer, while PreScission Protease, a GST fusion protein, remained bound to the resin. PreScission protease cleavage of GST–AT-3 fusion protein released an AT-3 form carry- ing five residues (GPLGS) upstream from the authentic N-terminus. Amino acid numbering, where presented, starts from the N-terminal methionine of authentic AT-3-Q6. Incubation of AT-3-Q6 and its mutants To monitor the appearance of proteolytic fragments, wild- type AT-3-Q6, as well as C14A, H119L and C14AH119L mutants, was incubated at 24 °C for different times in the presence of 50 mm Tris ⁄ HCl, pH 7.0, 150 mm NaCl, 1 mm EDTA, and 1 mm dithiothreitol (cleavage buffer). Each protein had a final concentration in the mixture of about 0.4 mgÆmL )1 . Fifteen-microliter samples were taken at dif- ferent times and subjected to SDS ⁄ PAGE. P. L. Mauri et al. Autolytic cleavage of ataxin-3 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS 4283 SDS ⁄ PAGE and western blotting SDS ⁄ PAGE was carried out according to Laemmli [38], in a Mighty Small apparatus (Hoefer Scientific Instruments, San Francisco, CA) using a 12% running gel and a 4% stacking gel. Proteins were revealed by Gel Code staining (Pierce Biotechnology, Rockford, IL). CD spectroscopy All far-UV CD spectra were collected at 25 °C using a Jasco-600 spectrophotometer (Jasco, Tokyo, Japan), and a cuvette with 0.1-cm pathlength. All experiments were per- formed in 50 mm Tris ⁄ HCl, pH 7.0, at a protein concen- tration of 0.1 mgÆmL )1 . The spectra were registered from 198 to 250 nm, and ran at a scan speed of 10 nmÆmin )1 , with a time response of 4 s and a data pitch of 0.2 nm. All the spectra were baseline-corrected and smoothed. The molar mean residue ellipticity [h] was expressed in degree- sÆcm 2 Ædmol )1 , and calculated as ½h¼h obs MWR=ð10lcÞ where h obs is the observed ellipticity in degrees, MWR the mean residue molar weight of the protein (121.1 mgÆ mmol )1 ), l the optical path length in centimeters, and c the protein concentration in grams per milliliter. Tryptic fragmentation of AT-3-Q6 in solution Sequence-grade modified trypsin (Promega, Madison, WI) was added to 50 lL of conditioned medium containing 4 lg of protein at a 1 : 50 enzyme ⁄ protein ratio (wt ⁄ wt) in 100 mm ammonium bicarbonate, pH 8.9, and incubated at 37 °C. Following an overnight incubation, the pH was adjusted to 2.0 by adding trifluoroacetic acid to stop the reaction. Ten microliters of the peptide mixture diluted 1 : 10 were subjected to LC ⁄ MS ⁄ MS. Trypsin digestion of AT-3-Q6 in the electrophoretic gel Protein bands were excised from the gel, cut into small fragments ( 1 · 1 mm) using a scalpel, and placed in a 1.5 mL siliconized tube. After dehydration in 100% meth- anol for 5 min at room temperature, fragments were rehydrated in 30% methanol for 5 min, and washed twice in ultrapure water for 10 min. The gel bands were washed three times for 10 min with 100 mm ammonium bicarbonate, pH 7.9, containing 30% acetonitrile, and fur- ther washed in ultrapure water. Gel fragments were thor- oughly dried and resuspended in 50 mm ammonium bicarbonate. Then, trypsin digestion was performed as previously described [39]. Trypsin was added at an enzyme ⁄ substrate ratio of 1 : 30 (wt ⁄ wt) in 100 mm ammonium bicarbonate, pH 7.9, and incubated over- night at 37 °C. The peptides were extracted from the gel at room temperature with 50% acetonitrile containing 0.1% trifluoroacetic acid. Aliquots containing 10 lLof sample were injected into the LC-MS ⁄ MS mass spectro- meter. LC conditions for LC-MS analysis For the analysis of enzymatic digests, a Phoenix 40 HPLC (Thermo Electron Corp., Milan, Italy) equipped with 7725i Rheodyne injector was coupled to an LCQ- Deca ion trap mass spectrometer by an electrospray inter- face. A Nucleosil C 18 column (0.5 · 150 mm, 5 lm; Waters, Milford, MA) with an acetonitrile gradient was used (eluent A, 0.1% formic acid in water; eluent B, 0.1% formic acid in acetonitrile). The flow rate was 25 lLÆmin )1 . The gradient profile was 10% B for 3 min, followed by 10–80% B for 60 min. For the analysis of intact AT-3-Q6 and polypeptides resulting from autolysis, aC 8 column (150 · 1 mm, 5 lm; Luna, Phenomenex, Torrance, CA) was used with a flow rate of 20 lLÆmin )1 . The gradient profile was 10% B for 3 min, followed by 10–80% B for 40 min. MS conditions The heated capillary was held at 260 °C at a voltage of 30 V. The spray voltage was 4.5 kV. For peptide analysis, spectra were acquired in automated MS ⁄ MS mode: each MS full scan (in the range 300–1600 m ⁄ z) was followed by three MS ⁄ MS scans of the most abundant ions, using a rel- ative collision energy of 35%. For protein and polypeptide analysis, MS full scan (in the range of 400–1800 m ⁄ z)in positive mode was used. Data handling of MS results Computer analysis of peptide MS ⁄ MS spectra was per- formed using Bioworks 3.1 SR1, based on the sequest algorithm (University of Washington, USA, licensed to ThermoFinnigan Corp., Austin, TX). The ‘no enzyme’ option was used. The experimental mass spectra produced were correlated to peptide sequences obtained by compar- ison with theoretical mass spectra of the Machado–Joseph disease protein 1 (AT-3) database downloaded from the Swiss-Prot website (http://www.expasy.org) (Q9CVD2). For peptide matching, the following limits were used: Xcorr scores greater than 1.5 for singly charged peptide ions, and 2.0 and 2.5 for doubly and triply charged ions, respectively. Data handling of MS spectra from LC-MS analysis of intact AT-3-Q6 and related polypeptides was performed with suitable tools for peptide fragment characterization in bioworks 3.1 SR1. Autolytic cleavage of ataxin-3 P. L. Mauri et al. 4284 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS Acknowledgements We are indebted to Gabriella Tedeschi for the N-ter- minal sequencing of AT-3 fragments. The excellent technical assistance of Enrico Rosti is gratefully acknowledged. This research was supported by grants from the Ministero dell’Universita ` e della Ricerca Sci- entifica e Tecnologica (MIUR, Rome, Italy) and from the Fondazione Cariplo (NOBEL project). References 1 Zoghbi HY & Orr HT (2000) Glutamine repeats and neurodegeneration. Annu Rev Neurosci 23, 217–247. 2 Cummings CJ & Zoghbi HY (2000) Trinucleotide repeats: mechanisms and pathophysiology. Annu Rev Genomics Hum Genet 1, 281–328. 3 Paulson HL, Das SS, Crino PB, Perez MK, Patel SC, Gotsdiner D, Fischbeck KH & Pittman RN (1997) Machado–Joseph disease gene product is a cytoplasmic protein widely expressed in brain. Ann Neurol 41, 453– 462. 4 Ross CA, Poirier MA, Wanker EE & Amzel M (2003) Polyglutamine fibrillogenesis: the pathway unfolds. Proc Natl Acad Sci USA 100, 1–3. 5 Clarke G, Collins RA, Leavitt BR, Andrews DF, Hay- den MR, Lumsden CJ & McInnes RR (2000) A one-hit model of cell death in inherited neuronal degenerations. Nature 406, 195–199. 6 Perutz MF & Windle AH (2001) Cause of neural death in neurodegenerative diseases attributable to expansion of glutamine repeats. Nature 412, 143–144. 7 Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli L, Zurdo J, Taddei N, Ramponi G, Dobson CM & Ste- fani M (2002) Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature 416, 507–511. 8 Zuccato C, Tartari M, Crotti A, Goffredo D, Valenza M, Conti L, Cataudella T, Leavitt BR, Hayden MR, Timmusk T et al. (2003) Huntingtin interacts with REST ⁄ NRSF to modulate the transcription of NRSE-controlled neuronal genes. Nat Genet 35, 76–83. 9 Zuccato C, Ciammola A, Rigamonti D, Leavitt BR, Goffredo D, Conti L, MacDonald ME, Friedlander RM, Silani V, Hayden MR et al. (2001) Loss of hunt- ingtin-mediated BDNF gene transcription in Hunting- ton’s disease. Science 293 , 493–498. 10 Albrecht M, Hoffmann D, Evert BO, Schmitt I, Wull- ner U & Lengauer T (2003) Structural modeling of ataxin-3 reveals distant homology to adaptins. Proteins 50, 355–370. 11 Masino L, Musi V, Menon RP, Fusi P, Kelly G, Frenkiel TA, Trottier Y & Pastore A (2003) Domain architecture of the polyglutamine protein ataxin-3: a globular domain followed by a flexible tail. FEBS Lett 549, 21–25. 12 Albrecht M, Golatta M, Wullner U & Lengauer T (2004) Structural and functional analysis of ataxin-2 and ataxin-3. Eur J Biochem 271, 3155–3170. 13 Chow MK, Mackay JP, Whisstock JC, Scanlon MJ & Bottomley SP (2004) Structural and functional analysis of the Josephin domain of the polyglutamine protein ataxin-3. Biochem Biophys Res Commun 322, 387–394. 14 Li F, MacFarlan T, Pittman RN & Chakravarti D (2002) Ataxin-3 is a histone-binding protein with two independent transcriptional corepressor activities. J Biol Chem 277, 45004–45012. 15 Wang G, Sawai N, Kotliarova S, Kanazawa I & Nukina N (2000) Ataxin-3, the MJD1 gene product, interacts with the two human homologs of yeast DNA repair protein RAD23, HHR23A and HHR23B. Hum Mol Genet 9, 1795–1803. 16 Madura K (2002) The ubiquitin-associated (UBA) domain: on the path from prudence to prurience. Cell Cycle 1, 235–244. 17 Hofmann K & Falquet L (2001) A ubiquitin-interacting motif conserved in components of the proteasomal and lysosomal protein degradation systems. Trends Biochem Sci 26, 347–350. 18 Donaldson KM, Li W, Ching KA, Batalov S, Tsai CC & Joazeiro CA (2003) Ubiquitin-mediated sequestration of normal cellular proteins into polyglutamine aggre- gates. Proc Natl Acad Sci USA 100, 8892–8897. 19 Chai Y, Berke SS, Cohen RE & Paulson HL (2004) Poly-ubiquitin binding by the polyglutamine disease protein ataxin-3 links its normal function to protein sur- veillance pathways. J Biol Chem 279, 3605–3611. 20 Scheel H, Tomiuk S & Hofmann K (2003) Elucidation of ataxin-3 and ataxin-7 function by integrative bio- informatics. Hum Mol Genet 12, 2845–2852. 21 Burnett B, Li F & Pittman RN (2003) The polygluta- mine neurodegenerative protein ataxin-3 binds polyubi- quitylated proteins and has ubiquitin protease activity. Hum Mol Genet 12, 3195–3205. 22 Berke SJ, Schmied FA, Brunt ER, Ellerby LM & Paul- son HL (2004) Caspase-mediated proteolysis of the polyglutamine disease protein ataxin-3. J Neurochem 89, 908–918. 23 McDonald WH, Ohi R, Miyamoto DT, Mitchison TJ & Yates JR3rd (2002) Comparison of three directly coupled HPLC MS ⁄ MS strategies for identification of proteins from complex mixtures: single-dimension LC-MS ⁄ MS, 2-phase MudPIT, and 3-phase MudPIT. Int J Mass Spectrom 219, 245–251. 24 Ghislain M, Dohmen RJ, Levy F & Varshavsky A (1996) Cdc48p interacts with Ufd3p, a WD repeat pro- tein required for ubiquitin-mediated proteolysis in Saccharomyces cerevisiae. EMBO J 15, 4884–4899. P. L. Mauri et al. Autolytic cleavage of ataxin-3 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS 4285 25 Doss-Pepe EW, Stenroos ES, Johnson WG & Madura K (2003) Ataxin-3 interactions with rad23 and valosin- containing protein and its associations with ubiquitin chains and the proteasome are consistent with a role in ubiquitin-mediated proteolysis. Mol Cell Biol 23, 6469– 6483. 26 Matsumoto M, Yada M, Hatakeyama S, Ishimoto H, Tanimura T, Tsuji S, Kakizuka A, Kitagawa M & Nakayama KI (2004) Molecular clearance of ataxin-3 is regulated by a mammalian E4. EMBO J 23, 659–669. 27 Polga ´ r L (2004) Catalytic mechanisms of cysteine pep- tidases. In Handbook of Proteolytic Enzymes, 2nd edn (Barrett AJ, Rawlings ND & Woessner F, eds), pp. 1072–1079. Elsevier, Academic Press, London. 28 Tarlac V & Storey E (2003) Role of proteolysis in poly- glutamine disorders. J Neurosci Res 74, 406–416. 29 Paulson HL (2000) Toward an understanding of poly- glutamine neurodegeneration. Brain Pathol 10, 293– 299. 30 Wellington CL, Ellerby LM, Hackam AS, Margolis RL, Trifiro MA, Singaraja R, McCutcheon K, Salvesen GS, Propp SS, Bromm M et al. (1998) Caspase cleavage of gene products associated with triplet expansion disor- ders generates truncated fragments containing the poly- glutamine tract. J Biol Chem 273, 9158–9167. 31 Wellington CL, Singaraja R, Ellerby L, Savill J, Roy S, Leavitt B, Cattaneo E, Hackam A, Sharp A, Thornber- ry N et al. (2000) Inhibiting caspase cleavage of hunt- ingtin reduces toxicity and aggregate formation in neuronal and nonneuronal cells. J Biol Chem 275, 19831–19838. 32 Sanchez I, Mahlke C & Yuan J (2003) Pivotal role of oligomerization in expanded polyglutamine neurodegen- erative disorders. Nature 421, 373–379. 33 Morley JF, Brignull HR, Weyers JJ & Morimoto RI (2002) The threshold for polyglutamine-expansion pro- tein aggregation and cellular toxicity is dynamic and influenced by aging in Caenorhabditis elegans. Proc Natl Acad Sci USA 99, 10417–10422. 34 Haacke A, Broadley SA, Boteva R, Tzvetkov N, Hartl FU & Breuer P (2006) Proteolytic cleavage of polygluta- mine-expanded ataxin-3 is critical for aggregation and sequestration of non-expanded ataxin-3. Hum Mol Genet 15, 555–568. 35 Goti D, Katzen SM, Mez J, Kurtis N, Kiluk J, Ben Haiem L, Jenkins NA, Copeland NG, Kakizuka A, Sharp AH et al. (2004) A mutant ataxin-3 putative-clea- vage fragment in brains of Machado–Joseph disease patients and transgenic mice is cytotoxic above a critical concentration. J Neurosci 24, 10266–10279. 36 Nicastro G, Menon RP, Masino L, Knowles PP, McDonald MO & Pastore A (2005) The solution struc- ture of the Josephin domain of ataxin-3: structural determinants for molecular recognition. Proc Natl Acad Sci USA 102, 10493–10498. 37 Shehi E, Fusi P, Secundo F, Pozzuolo S, Bairati A & Tortora P (2003) Temperature-dependent, irreversible formation of amyloid fibrils by a soluble human ataxin- 3 carrying a moderately expanded polyglutamine stretch (Q36). Biochemistry 42, 14626–14632. 38 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 39 Forler D, Kocher T, Rode M, Gentzel M, Izaurralde E & Wilm M (2003) An efficient protein complex purifica- tion method for functional proteomics in higher eukar- yotes. Nat Biotechnol 21, 89–92. Autolytic cleavage of ataxin-3 P. L. Mauri et al. 4286 FEBS Journal 273 (2006) 4277–4286 ª 2006 The Authors Journal compilation ª 2006 FEBS . Ataxin-3 is subject to autolytic cleavage Pier Luigi Mauri 1 , Matteo Riva 2 , Daniela Ambu 1 , Antonella De Palma 1 , Francesco Secundo 3 , Louise Benazzi 1 , Marco Valtorta 2 , Paolo Tortora 2, *. conclude that AT-3 activity is not caspase-like. The ability of AT-3 to undergo autolytic degradation might be related to the mechanisms of pathogenesis, as proteolysis is widely thought to play a role in. consistently observed that pure AT-3 undergoes proteolysis when incubated at room temperature for several hours, and this prompted us to check whether this phenomenon is supported by autolytic cleavage. Based