Proteasomal Degradation and N-terminal Protease Resistance of the Codon 145 Mutant Prion Protein*

An amber mutation at codon 145 (Y145stop) of the prion protein gene results in a variant of an inherited human prion disease named Gerstmann-Sträussler-Scheinker syndrome. The characteristic features of this disorder include amyloid deposits of prion protein in cerebral parenchyma and vessels. We have studied the biosynthesis and processing of the prion protein containing the Y145stop mutation (PrP145) in transfected human neuroblastoma cells in an attempt to clarify the effect of the mutation on the metabolism of PrP145 and to gain insight into the underlying pathogenetic mechanism. Our results demonstrate that 1) a significant proportion of PrP145 is not processed post-translationally and retains the N-terminal signal peptide, 2) most PrP145 is degraded very rapidly by the proteasome-mediated pathway, 3) blockage of proteasomal degradation results in intracellular accumulation of PrP145, 4) most of the accumulated PrP145 is detergent-insoluble, and both the detergent-soluble and -insoluble fractions are resistant to mild proteinase K (PK) treatment, suggesting that PK resistance is not simply because of aggregation. The present study demonstrates for the first time that a mutant prion protein is degraded through the proteasomal pathway and acquires PK-resistance if degradation is impaired.

dation, and in some cases, can transmit the disease (PrP Res ) (1)(2)(3). Recent NMR studies on recombinant PrP have shown that the N-terminal domain (23-120 in mouse recombinant PrP) is highly flexible and has a random coil structure, whereas the C-terminal region (129 -219) contains two short ␤-sheet structures and three ␣-helical domains (4,5). The major conformational change that causes PrP C to become the pathogenic and infectious PrP Res isoform is thought to involve refolding of the region between residues 90 and 112, which would lead to conversion of the region containing the two short ␤-sheet structures and of the first ␣-helix into a large ␤-sheet formation (3,6). However, the remaining C-terminal structures including the two other ␣-helices and the disulfide bond need to be preserved for PrP Res to be infectious (3,6).
The events leading to the conversion of PrP C to PrP Res are, at present, not fully understood. In the inherited prion diseases, which have only been associated with mutations in the PrP gene (PRNP), the mutation is believed to destabilize the mutant PrP (PrP M ), which then undergoes a spontaneous conformational change into the protease-resistant and pathogenic form (7)(8)(9). Twenty-three pathogenic mutations in PRNP have been reported to date, which are associated with three phenotypes: Creutzfeldt-Jakob disease, fatal familial insomnia, and Gerstmann-Strä ussler-Scheinker disease (GSS) (8,9). Despite their congenital presence, all mutations cause diseases that become symptomatic in the adult or advanced age. Of the seven PRNP mutations associated with GSS, a chronic cerebellar ataxia and dementia characterized by the presence of prominent amyloid plaques containing internal PrP M fragments, all are missense mutations except for the mutation at codon 145 that replaces tyrosine (TAT) with a stop (TAG) codon (Y145stop) (10 -12).
The 145 mutation is especially challenging because it results in the premature termination of protein synthesis and yields a truncated PrP (PrP 145 ) that lacks the C-terminal 146 -231 amino acids including the glycosylphosphatidyl inositol anchor, which links PrP to the cell surface, and the two sites for N-glycosylation, which are known to stabilize the PrP molecule (13,14). Thus, although PrP 145 includes the 90 -112 segment where the major conformational changes take place and almost all the ϳ80 -147 internal fragment that is found in PrP amyloid, it lacks the C-terminal region containing the two ␣-helices and the disulfide bond, which are required for the PrP C -PrP Res conversion (15,16). Moreover, most of the PrP 145 is likely to include the 23-90-residue N-terminal region of PrP, which has been shown to be superfluous for conversion to PrP Res (15)(16)(17). Yet, the Y145stop mutation is associated with a phenotype not basically different from that of the GSS subtypes associated with other mutations, which do not result in the truncation of PrP M (10). The only significant difference is the presence of numerous PrP-amyloid deposits in the cerebral vessels rather than in the brain parenchyma as in the other GSS subtypes (10 -12).
Currently, there are no animal or cellular models of the Y145stop mutation. Expression of the truncated PrP 145 was not detected in transgenic animals and transfected neuroblastoma cells following deletion of the 144 -231 region, which results in a PrP M identical in primary structure to that generated by the Y145stop mutation (17,18). To examine the effects of this mutation on the metabolism of PrP 145 , hence to gain a better understanding of the mechanisms involved in the pathogenesis of this GSS variant, we have transfected human neuroblastoma cells with PRNP constructs carrying the Y145stop mutation or wild type PRNP. We report, for the first time, that mutant PrP 145 is degraded through the proteasomal pathway. Inhibition of proteasomal degradation results in the accumulation of PrP 145 in intracellular compartments, including the endoplasmic reticulum (ER), the cis-medial-Golgi compartment, and the nucleus. Most of the accumulated PrP 145 is aggregated and partially resistant to mild proteinase K treatment. Proteaseresistant PrP 145 is also present in the detergent-soluble fraction, suggesting that the PrP 145 protease resistance is not simply because of aggregation.

Materials, Cell Culture Conditions, and Production of Transfected
Cell Lines-Opti-MEM, fetal bovine serum, penicillin/streptomycin, methionine, and cysteine-free Dulbecco's modified Eagle's medium, and Lipofectin were from Life Technologies Inc.; hygromycin B and lactacystin were from Calbiochem; Tran 35 S-label was from ICN; protein A-Sepharose was from Amersham Pharmacia Biotech; all other chemicals were purchased from Sigma. Transfected M17 cells expressing wild type or mutant prion protein (Y145amber) were generated as described (13,19). All cultures were maintained at 37°C in Opti-MEM supplemented with 5% fetal calf serum and penicillin-streptomycin in a humidified atmosphere containing 5% CO 2 . Cultures of transfected cells were supplemented with 500 g/ml hygromycin. The following antibodies were used: anti-N, rabbit antiserum to synthetic peptide corresponding to human PrP residues 23-40 (B. Ghetti, Indiana University); 3F4, a monoclonal antibody that recognizes an epitope on human PrP residues 109 -112 (R. Kascsak, New York State Institute for Basic Research in Developmental Disabilities); anti-protein disulfide isomerase (M. Lamm, Case Western Reserve University), anti-calnexin rabbit immune serum (A. Helenius, Yale University), anti-␣-mannosidase II (M. Farquhar, Univ. California, San-Diego), anti-cathepsin-D (R.A Nixon, Harvard University), and anti-Man-6-phosphate (L. Traub, Washington University School of medicine).
In Vitro Transcription and Translation-The codon 145 mutant was originally produced in the phagemid pVZ1, which contains both T7 and SP6 bacteriophage RNA polymerase promoters, using oligonucleotidedirected mutagenesis. Using a BamHI-cleaved template, run-on capped RNA was produced using the Cap-scribe system (Roche Molecular Biochemicals) as recommended by the manufacturer. Transcripts were analyzed using ethidium bromide-stained gels to assess their purity. The in vitro transcription products were translated into protein using the Promega message-dependent rabbit reticulocyte system with or without added canine pancreatic microsomes (Promega) to cleave the signal peptide. To control for microsome activity, a transcript derived from ␤-lactamase was used. The conditions used in the translation reaction were essentially as described by the manufacturer.
Metabolic Labeling, Immunoprecipitation, and Western Blots-In a typical experiment, 9 ϫ 10 6 cells were used for each condition. Equal amounts of total protein was used from cells expressing either normal or mutant PrP. Immunoprecipitation and Western blots were performed essentially as described (19), with the following modifications. For pulse-chase experiments, cells were preincubated in the presence or absence of the indicated inhibitors (lactacystin 80 M, N-acetyl-leucylleucyl-norleucinal (ALLN) 80 M, brefeldin A 5 g/ml) for 1 h before labeling with 0.166 mCi/ml of Tran 35 S-label (ICN) in methionine-cysteine-free Dulbecco's modified Eagle's medium with 5% dialyzed serum. Where indicated, appropriate inhibitors were included in the labeling and chase medium. At the indicated times, the medium was collected to check any secreted PrP 145 , and the cells were lysed in a buffer containing 0.5% Nonidet P-40, 0.5% deoxycholate, and 10 mM EDTA in Trisbuffered saline (20 mM Tris, 150 mM NaCl, pH 7.4) containing a mixture of protease inhibitors. Cell debris was cleared by centrifugation at 290 ϫ g, and the clarified cell lysate and medium samples were subjected to immunoprecipitation with the appropriate antibodies in the presence of 1% bovine serum albumin and 0.1% N-lauryl sarcosine. Protein-antibody complexes bound to protein-A-Sepharose (Amersham Pharmacia Biotech) were washed four times with 0.5 ml of wash buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.8, 0.1% N-lauryl sarcosine, and 0.1 mM phenylmethylsulfonyl fluoride), and bound protein was eluted by boiling in sample buffer (Tris-HCL, pH 6.8, 3% SDS, 10% glycerol, 5% ␤-mercaptoethanol) and analyzed by SDS-PAGE fluorography. PrP bands were quantitated by PhosphorImager analysis (Molecular Dynamics). For Western blots, proteins from cell lysates were precipitated with 5 volumes of cold methanol at Ϫ20°C, fractionated by SDS-PAGE, and electrophoretically transferred to Immobilon-P (Millipore) for 2.5 h at 70 volts at 4°C. Membranes containing transferred proteins were blocked in Tris-buffered saline containing 10% nonfat dry milk and 0.1% Tween 20 for 1 h at 37°C and probed with anti-PrP antibodies (anti-N diluted 1:4000, 3F4 diluted 1:50,000, or anti-C diluted 1:3000) dissolved in antibody dilution buffer (Tris-buffered saline, 1% normal goat serum, and 0.05% bovine serum albumin). Immunoreactive bands were detected with the appropriate secondary antibody conjugated to horseradish peroxidase (anti-rabbit diluted 1:3000, anti-mouse diluted 1:4000) and visualized on an autoradiographic film by ECL (Amersham Pharmacia Biotech). To quantitate the relative density of immunoreactive bands, exposed autoradiographic film was scanned at 42-mm resolution with a GE10 densitometer and quantitatively analyzed using Quantity One software (PDIG20, QS30).
For biotinylation of surface proteins, untreated, or cells treated with ALLN for 2 h were biotinylated with 0.2 mg/ml of sulfo-N-hydroxysuccinimide-biotin in PBS for 15 min on ice. Excess biotin was quenched with 50 mM glycine (in PBS), and after three more washes with PBS, the cells were lysed as above and subjected to immunoprecipitation with 3F4. The immunoprecipitated proteins were fractionated by SDS-PAGE and electroblotted to Immobilon-P, and the biotinylated PrP was detected by horseradish peroxidase-conjugated streptavidin and ECL.
Detection of Associated Chaperone Proteins with PrP 145 -Cells expressing PrP C or PrP 145 were radiolabeled for 2 h in the presence or absence of 80 M ALLN and lysed with a nondenaturing buffer containing 1% CHAPS or 2% Triton X-100 in the presence of a mixture of protease inhibitors. The lysate was subjected to immunoprecipitation with anti-KDEL (Stressgen), anti-calnexin, or anti-Grp 94 antibodies (Stressgen) and fractionated by SDS-PAGE to check co-immunoprecipitation of any of the PrP C or PrP 145 forms with the above ER chaperones. In a parallel experiment, the PrP was immunoprecipitated from the lysates, and the presence of any associated chaperones was evaluated by immunoblotting the electrophoretically transferred proteins by specific antibodies.
Detection of Ubiquitinated PrP-For detecting ubiquitinated PrP 145 , untreated and ALLN-treated PrP C -and PrP 145 -expressing cell lysates were immunoprecipitated with 3F4 as above. After fractionating on SDS-PAGE, the immunoprecipitated proteins were transblotted and probed with anti-ubiquitin antibodies 1510, 5-25 (Chemicon), or LB112 (J.Q. Trojanowski, University of Pennsylvania). Immunoreactive bands were detected with appropriate horseradish peroxidase-conjugated secondary antibodies. Alternatively, PrP C and PrP 145 cells were transfected with myc-tagged ubiquitin-expressing plasmids H 6 M-Ub, or H 6 M-Ub K48R (dominant negative ubiquitin mutant) (Ron R. Kopito, Stanford University). Steady state levels of PrP C and PrP 145 were determined in transient transfectants by Western blotting. To detect ubiquitinated PrP 145 , transient transfectants were radiolabeled in the presence of ALLN, immunoprecipitated with anti-myc antibody to isolate myc-tagged ubiquitin-conjugated PrP 145 , and analyzed by SDS-PAGE.
Confocal Immunofluorescence Microscopy-For immunofluorescent staining of PrP, transfected cells expressing PrP C or PrP 145 were grown on poly-D-lysine-coated glass coverslips overnight and treated with lactacystin in complete medium for the indicated times. Control cells (0 h) received only medium. For washout experiments, cells treated with lactacystin for 4 h were washed with complete medium, and incubated in fresh medium for 0 -4 h. At the indicated times, cells were rinsed with PBS and fixed in 3% paraformaldehyde for 30 min at room temperature. Free aldehyde groups were quenched with 50 mM NH 4 Cl (in PBS), and the cells were permeabilized with PBS containing 0.1% Triton X-100, 0.1% SDS, 2.5 mM MgCl 2 , and 5 mM KCl for 5 min. Nonspecific sites were blocked with PBS containing 10% goat serum and 0.2% bovine serum albumin, followed by 0.2% gelatin in PBS. Cells were then incubated with anti-PrP antibody (3F4, diluted 1:25), followed by fluorescein-conjugated secondary antibody for 35-40 min each. For subcellular immunolocalization, subsequent incubations were done with anti-calnexin, anti-␣-mannosidase II, anti-cathepsin-D, or anti-mannose-6-phosphate antibodies, respectively, followed by rhodamine-conjugated secondary antibodies. The cells were rinsed in PBS, mounted in gel-mount (Biomeda Corp., Fostar City, CA), and observed using a laser scanning confocal microscope (Bio-Rad). To compare the fluorescence intensity of lactacystin-treated and washout samples accurately, all experimental conditions were kept constant, including antibody titer, total cell confluence, laser strength, magnification, and basal instrument settings. A single 0.5-m optical section was photographed in each case.
Assay of "Detergent Insolubility" and Proteinase K Resistance-Aggregation, partial resistance to proteinase K (PK) digestion, and insolubility in nonionic detergents is the hallmark of PrP Res . However, when applied in this context, detergent insolubility implies insolubility of PrP in a buffer containing 0.5% Nonidet P-40 and 0.5% sodium deoxycholate when centrifuged at 100,000 ϫ g for 1 h. Untreated, and cells treated with lactacystin for two h were lysed at 4°C in 1 ml of Tris buffer containing 0.5% of Nonidet P-40 and sodium deoxycholate each. After a brief centrifugation at 290 ϫ g, cell debris (P 1 ) was discarded, and the supernatant (S 1 ) was divided into two equal parts. One part was set aside, and the other was centrifuged at 100,000 ϫ g for 1 h at 4°C to obtain a high speed supernatant (S 2 ) and pellet fraction (P 2 ). The pellet fraction P 2 was resuspended in the same volume of buffer as S 2 and sonicated. Equal aliquots of S 1 (before ultracentrifugation), S 2 , and P 2 fractions were immunoblotted with 3F4. The S 2 and P 2 fractions of lactacystin-treated cells were also assayed for resistance to mild PK digestion as described previously (19).
To check if aggregated PrP 145 is immunoprecipitated efficiently, lysates from cells incubated in the presence or absence of 80 M of ALLN for 2 h were subjected to ultracentrifugation as described above to separate the S 1 , S 2, and P 2 fractions. Each fraction was immunoprecipitated with 3F4, and the remaining supernatant was methanol-precipitated and immunoblotted with 3F4 to detect any nonimmunoprecipitated PrP remaining in the supernatant. The 15.5-kDa band was consistently observed in the supernatant of ALLN-treated cells. However, if the lysates were first boiled in the presence of 0.5% SDS, diluted 5-fold, and then subjected to immunoprecipitation, no PrP was detected in the supernatant, suggesting that the inefficient immunoprecipitation of the 15.5-kDa form is because of aggregation.

PrP 145 Is Expressed as Two Isoforms at Steady State-On
blots immunostained with the anti-PrP antibody 3F4 (to residues 109 -112), the normal or cellular prion protein (PrP C ) migrates as three bands corresponding to the unglycosylated (U), intermediate (I), and highly glycosylated (H) forms of 27, 29 -30, and ϳ33-42 kDa, respectively (Fig. 1A). In contrast, PrP 145 migrates as two bands, a lower band of 14 kDa (PrP14), the expected molecular mass for PrP with the 145 amber mutation, and an upper, more prominent band of 15.5 kDa (PrP15.5) that accounts for 66% of the total PrP 145 (Fig. 1A). As expected, both bands are readily detected with the 3F4 and anti-N antibodies but not with anti-C-terminal antibody (data not shown). Immunoprecipitation of PrP 145 (which includes both PrP14 and PrP15.5) from cells radiolabeled with [ 35 S]methionine and cysteine for 2 h with the 3F4 antibody shows a similar pattern (Fig. 1B), except that PrP15.5 accounts for only 25% of the total PrP 145 . The cell-associated pool of PrP 145 at steady state is approximately 9 times smaller than that of PrP C (Fig. 1A). Moreover, Ͻ1% of the PrP 145 is recovered from the medium, suggesting that the low expression of PrP 145 is not because of secretion.
Three experiments were performed to determine whether PrP15.5 represents a form of PrP 145 with an uncleaved Nterminal signal peptide of 22 amino acids: 1) metabolic labeling with [ 35 S]cysteine, a residue that is present only in the signal peptide; 2) cell-free translation with radiolabeled methionine and cysteine or only cysteine in the absence or presence of microsomes to obtain a translation product with or without the signal peptide, respectively; 3) a short pulse of 30 s with 35 Slabeled methionine and cysteine followed by a chase to inves-tigate whether PrP15.5 converts to the PrP14 form. Metabolic labeling of cells with cysteine for 2 h shows only PrP15.5, whereas both forms are detected when cells are radiolabeled with methionine and cysteine (data not shown). After cell-free translation with radiolabeled methionine and cysteine in the absence of microsomes, only PrP15.5 is retrieved, most of which is converted to the PrP14 form when the microsomes are added co-translationally (Fig. 1C). Translation in the presence of radiolabeled cysteine yields only PrP15.5, which disappears when microsomes are added, confirming that PrP15.5 includes the N-terminal signal peptide that is lost on addition of microsomes (Fig. 1C). The bands obtained in vitro co-migrate with PrP15.5 and PrP14 from radiolabeled cells (Fig. 1C). The short pulse-chase experiment shows that at the end of a 30-s pulse, PrP15.5 is predominant and accounts for 55% of the total PrP 145 (Fig. 1D). The ratio between the two forms is reversed at the end of 2.5 min (Fig. 1D), whereas the total PrP 145 remains unchanged during this time period. This finding suggests that PrP15.5 is converted into the PrP14 form. Both forms then decrease rapidly during the chase, so that only 16% of total PrP 145 remains after 30 min, most of which consists of PrP14 (Fig. 1D).
Immunofluorescence analysis by double immunostaining with anti-PrP and an antibody to calnexin, an endoplasmic reticulum (ER)-specific protein, shows that although PrP C is mostly distributed at the cell surface and in the Golgi region (Fig. 1E, panel 1) (19), a very small amount of PrP 145 is detected in an intracellular compartment, and it co-stains with the Golgi marker ␣-mannosidase II (Fig. 1E, panel 3). No ER localization is observed with the ER marker calnexin (Fig. 1E, panel 2), probably because the small quantity of PrP 145 that escapes degradation transits through the ER very rapidly. No PrP 145 is detected on the plasma membrane either by immunostaining (Fig. 1E, panels 2 and 3) or cell surface biotinylation (data not shown), excluding the possibility that PrP15.5 is inserted into the cell membrane through the signal peptide.
Together, these results show that 1) PrP 145 is unstable and is not detected in the culture medium in significant amounts, 2) PrP15.5 represents a PrP 145 form with an uncleaved N-terminal signal peptide even though it is apparently translocated efficiently into the ER (see below), 3) no PrP 145 is detected at the cell surface. The very low expression of PrP 145 compared with PrP C suggests that PrP 145 turns over very rapidly. The under-representation of PrP15.5 after radiolabeling and immunoprecipitation (25%) when compared with the steady state level by Western blot analysis (66%) (Figs. 1, A versus B) raises the possibility that this form is not immunoprecipitated efficiently because of a change in its conformation, as previously observed with the PrP M Q217R (19) (see below).
PrP 145 Is Rapidly Degraded in a Pre-Golgi Compartment by the Proteasomal Pathway-To determine whether PrP 145 is degraded by the lysosomes or in a pre-Golgi compartment, we carried out pulse-chase analysis either in the presence of various lysosomal inhibitors (leupeptin, ammonium chloride, or chloroquine) or by blocking transport beyond the ER-cis-Golgi compartment by incubating cells at 15°C or treating them with brefeldin A (20). Neither the inhibition of lysosomal activity (data not shown) nor the block of vesicular transport at low temperature ( Fig. 2A) or brefeldin A (Fig. 2B) blocked the degradation of PrP 145 , although, as expected, the rate of degradation was slower at 15°C as compared with 37°C (see Fig.  1D). A similar analysis at 37°C shows that although PrP C matures into various glycoforms and is stable, 70% of the total being present after 2 h of chase, most of the PrP 145 disappears rapidly, and only 12% remains after 2 h (data not shown). Taken together, these results demonstrate that PrP 145 turns over in a pre-Golgi compartment and is not degraded through the lysosomal pathway.
Because both membrane and secretory proteins can be degraded through the proteasomal pathway (21-23), we evaluated this possibility by treating cells expressing PrP 145 with the proteasomal inhibitor lactacystin or ALLN during pulsechase experiments. In cells treated with lactacystin or ALLN, 48 and 36%, respectively, of PrP 145 remain after a 2-h chase, as compared with 12% in untreated cells ( Fig. 2C; p Ͻ3 ϫ 10 Ϫ4 at 1 h*, and p Ͻ 6 ϫ 10 Ϫ5 at 2 h**; n ϭ 3). PrP14 accounts for most of the protected PrP 145 (Fig. 2C). There is no change in the kinetics of turnover of PrP C in the presence of ALLN or lacta-cystin under the same experimental conditions (data not shown). To check if the PrP 145 that accumulates intracellularly following proteasomal inhibition is ubiquitinated, ALLNtreated cells were immunoprecipitated with 3F4, and the immunoprecipitates were immunoblotted with a panel of antiubiquitin antibodies to detect any ubiquitinated PrP forms. Alternately, PrP 145 cells were transfected with normal or a dominant negative mutant of ubiquitin followed by immunoprecipitation with 3F4. No ubiquitinated PrP 145 was detected in either case, and the co-expression of mutant ubiquitin did not stabilize PrP 145 (data not shown).
To evaluate if PrP 145 rescued from proteasomal degradation  1-6). E, immunofluorescent staining of PrP with 3F4 (green) and of calnexin, an ER marker (red), shows that PrP C is localized to the cell surface and the Golgi region (panel 1). Similar staining of PrP 145 -expressing cells shows a small amount of PrP 145 at steady state, mainly co-localizing with the Golgi marker ␣-mannosidase II (red; panel 3). No co-localization of PrP 145 is seen with the ER marker calnexin (red; panel 2).
is secreted, cells expressing PrP 145 were radiolabeled for 2 h in the continuous presence of lactacystin or ALLN, and the cell lysate and culture medium were subjected to immunoprecipitation with 3F4. Only PrP15.5 was secreted, although the intracellular pool of PrP14 exceeded that of PrP15.5 (Fig. 2D). No PrP 145 form was detected in the medium if cells were labeled in the presence of brefeldin A (data not shown).
Accumulated PrP 145 Is Aggregated and Less Sensitive to Protease Digestion-To investigate whether the PrP 145 that accumulates in the absence of proteasomal degradation is aggregated, untreated cells and cells treated with lactacystin for 2 h were lysed in a buffer containing the nonionic detergents Nonidet P-40 and sodium deoxycholate, conventionally used for detecting aggregated PrP. After pelleting the cell debris (P 1 ) by a low speed centrifugation, the supernatant (S 1 ) was centrifuged at 100,000 ϫ g for 1 h to separate a soluble (S 2 ) and an insoluble (P 2 ) fraction. In untreated cells, almost all of the PrP 145 is recovered in the high speed supernatant or soluble (S 2 ) fraction, and a small amount of PrP15.5 is detected in the P 2 fraction (Fig. 3A, upper panel). In lactacystin-treated cells, the amount of PrP 145 is at least 4-fold higher, and a significant proportion of it is present in the pellet (P 2 ) as an insoluble fraction. This fraction contains mostly PrP15.5 (Fig. 3A, lower  panel) and increases in amount with extended chase time (data not shown). The insolubility of PrP15.5 and inefficient immunoprecipitation especially in the presence of proteasomal inhibitors was further confirmed when we subjected untreated and lactacystin-treated cell lysates to immunoprecipitation and immunoblotted the remaining supernatant with the same antibody. In untreated cells, virtually no PrP14 or 15.5 could be detected in the supernatant, whereas in lactacystin-and ALLN-treated cells, significant amounts of mostly PrP15.5 remained in the supernatant with nonimmunoprecipitated proteins. However, when the lysates were boiled in the presence of SDS before immunoprecipitation, no PrP15.5 remained in the supernatant in treated or untreated cells (data not shown). These findings may explain the preferential detection of PrP15.5 in immunoblots as compared with pulse-labeled immunoprecipitates (see Figs. 1, A and B) if the presence of the signal peptide induces a change in its conformation that leads to inefficient immunoprecipitation. None of the PrP 145 forms (PrP14 or 15.5) were associated with any of the ER chaperones, Grp78, Grp94, or calnexin (data not shown).
The sensitivity to PK of the PrP 145 that accumulates follow- ing treatment with lactacystin was tested on immunoblots of lactacystin-treated cells by digesting the S 2 and P 2 fractions with 3.3 g/ml of PK for 1 to 10 min. PrP C (data not shown) and untreated PrP 145 were degraded completely after 1 min of PK treatment (Fig. 3B). In contrast, a 1-10-min treatment with PK of S 2 fraction obtained from mutant cells exposed to lactacystin resulted in the presence of two additional bands of ϳ14 and ϳ12 kDa. This finding suggests that both PrP14 and 15.5 comprise a 1.5-2 kDa smaller fragment, which is weakly resistant to PK treatment (Fig. 3B). However, the gel mobility of the "PK-resistant" PrP 145 forms makes the interpretation of the data difficult and the possibility that the two PK-resistant bands represent the intact and truncated forms of PrP14, whereas PrP15.5 is completely digested cannot be excluded. Paradoxically, PrP15.5 present in the insoluble P 2 fraction is less PK-resistant, because it is almost completely digested after a 5-min treatment (Fig. 3B). Nonetheless, after a 1-min treatment, a small amount of the 1.5-2-kDa lower band is still detectable, supporting the conclusion that PrP15.5 also generates a smaller PK-resistant fragment (Fig. 3B). The ϳ14-kDa and ϳ12-kDa PK-resistant PrP 145 forms were detected, not only by the 3F4 antibody but also by an antibody to the PrP N terminus (antiserum to residues 23-40 of PrP) (data not shown). Thus, the PK digestion must occur at the PrP 145 C terminus, beyond residue 112, which is recognized by 3F4, and not around residue 90, as observed in the full-length PrP Res . Overall, only ϳ1% of the PrP 145 remains after 10 min of PK digestion (Fig. 3B).
PrP 145 Rescued from Proteasomal Degradation Is Found in the Nucleus-The intracellular distribution of PrP 145 accumulated after inhibition of proteasomal function was analyzed by immunofluorescence of PrP 145 in cells treated with lactacystin for 0 -4 h (Fig. 4). In untreated cells, PrP 145 immunoreactivity was restricted to the Golgi region, as previously shown (Fig. 4A,  0 h, left panel; Fig. 1E, panel 3). After incubation with lactacystin for 1-4 h, the reactivity was detected in punctate and vesicular structures, although it remained most intense in the Golgi region (Fig. 4A, 1-4 h, left panels). Strikingly, a diffuse PrP staining of the nucleus sparing the nucleolus was also detected with longer treatment (Fig. 4A, 2-4 h, left panels). Replacement of the lactacystin-containing medium with normal medium followed by a chase for 4 h gradually restored the original PrP 145 distribution (Fig. 4A, right panels, 0 -4 h). PrP immunoreactivity first decreased in punctate and vesicular structures (Fig. 4A, 0 -1 h, right panels), then in the nucleus (Fig. 4A, 1-2 h, right panels), and finally was detected only in the Golgi region as in the untreated cells expressing PrP 145 (Fig. 4A, 4 h, right panel). The vesicular staining co-localized with calnexin, an ER marker (Fig. 4B, right panel), whereas the reactivity adjacent to the nucleus co-localized with the Golgi marker ␣-mannosidase II (Fig. 4B, left panel). Nuclear immunoreactivity of PrP 145 in lactacystin-treated cells co-localized with the nuclear marker DAPI and was also observed after immunostaining of deplasticized sections of plastic-embedded cell preparations as well as with immunofluorescence of isolated nuclear fractions (data not shown). No immunoreactivity was observed in the nuclei of PrP C -expressing or nontransfected M17 neuroblastoma cells treated with lactacystin or ALLN for 4 h (data not shown). The accumulated PrP 145 did not colocalize with the late endosomal-lysosomal marker cathepsin-D or with the lysosomal marker mannose 6-phosphate (data not shown). Thus, upon inhibition of proteasomal degradation, PrP 145 accumulates in membrane-bound compartments, including the ER, Golgi apparatus, and the nucleus. DISCUSSION The Y145stop mutation in the human prion protein gene, PRNP, is associated with a GSS variant of prion disease. Previous attempts to generate a model of this GSS variant in transgenic mice and transfected cells have failed because no expression of the mutant PrP 145 could be detected in these models (18,24). We now demonstrate that in a transfected cell model, PrP 145 is expressed in two truncated forms, one of which conserves the signal peptide. Both forms are unstable and are rapidly degraded through the proteasomal pathway. However, both accumulate in significant quantities in intracellular compartments and become aggregated and weakly protease-resistant when proteasomal degradation is impaired. These findings A, untreated cells and cells pretreated with lactacystin for 2 h were lysed with nonionic detergents, and cell debris was pelleted by a low speed centrifugation. The low speed supernatant or detergent-soluble (S 1 ) fraction was centrifuged at 100,000 ϫ g to obtain high speed detergent soluble (S 2 ) and insoluble (P 2 ) fractions. Immunoblotting of an aliquot of each fraction with 3F4 shows that most of the PrP 145 from untreated cells partitions in the S 2 fraction (lane 2), and very little is recovered in the pellet P 2 (lane 3). In lactacystin-treated cells, in contrast, most of the PrP 145 fractionates in the insoluble P 2 fraction, where PrP15.5 accounts for virtually all the PrP 145 (lane 3) (the amount of PrP 145 in lactacystin-treated lysates is at least 4-fold higher than in untreated cells. An underexposed fluorogram is shown for the sake of clarity). B, immunoblotting of PrP 145 present in the soluble (S 2 ) and insoluble (P 2 ) fractions obtained from cells exposed to lactacystin for 2 h reveals the presence of PK-resistant PrP 145 in both fractions after treatment with 3.3 g/ml of proteinase K. In S 2 , both PrP14 and 15.5 forms are present and appear to migrate ϳ1.5-2 kDa faster after PK treatment, perhaps because of the generation of a shorter PK-resistant fragment (lanes 3-6) (see text for alternative explanations). A significant amount of the PrP15.5 in the P 2 fraction also resists 1 min of PK treatment (lanes 7-10). In contrast, untreated PrP 145 is completely degraded after a 1-min treatment (lanes 1-2) (untreated PrP 145 in lanes 1 and 2 is overexposed to emphasize its complete digestion by PK in 1 min). may resolve the dilemma posed by the previous models. They also widen the spectrum of pathogenetic mechanisms that may be involved in prion diseases and provide novel avenues of investigation toward the understanding of this puzzling GSS variant.
The Uncleaved Signal Peptide Predisposes PrP 145 to Aggregation-Inefficient cleavage of the N-terminal signal peptide because of naturally occurring mutations within the signal has been shown to be pathogenic in various conditions, but it is unprecedented in prion diseases (25)(26)(27)(28). This "proform" appears to accumulate intracellularly and tends to aggregate more readily than the signal-cleaved form. The inefficient immunoprecipitation of this form is probably because of a change in conformation of its "soluble" pool, in addition to the formation of detergent-insoluble aggregates (see below). Thus, under normal experimental conditions, the amount of the signal peptide containing PrP 145 recovered in immunoblots is more than four times the amount recovered after immunoprecipitation. After lactacystin treatment, the signal peptide containing PrP 145 accounts for almost all of the detergent-insoluble and weakly protease-resistant aggregates that accumulate intracellularly. With continuous lactacystin treatment, both the sig-nal-uncleaved and -cleaved forms are secreted into the medium through a brefeldin A-sensitive pathway, although the signaluncleaved form comprises the major secreted form. The preferential detection of the latter in the medium could be because of its greater stability. Thus, both forms translocate into the ER lumen, and the signal-uncleaved form is not inserted in the lipid bilayer through the signal peptide. This conclusion is consistent with the lack of detectable PrP 145 on the cell surface by either immunofluorescence or biotinylation. None of the PrP 145 forms were found to be bound to any of the major ER-specific chaperones, either in the presence or absence of lactacystin.
PrP 145 Is Degraded by the Proteasome-The PrP 145 has a half-life of ϳ10 min and at steady state is nine times less abundant than PrP C . Therefore, it is by far the most unstable of all the forms of mutant PrP we have examined to date (13,19). The lack of all major post-translational modifications and presence of the signal peptide, both of which target PrP 145 for rapid degradation and aggregation, easily explain the marked instability of PrP 145 .
The turnover of both PrP 145 forms that persists at 15°C and in the presence of brefeldin A point to a pre-Golgi site of degradation. Following inhibition of proteasomal degradation, PrP 145 accumulates primarily in the ER, Golgi, and in the nucleus, but apparently not in the late endosomes or lysosomes. The precise site of PrP 145 proteasomal degradation has not been established in this study. PrP 145 might be degraded by proteasomes on the cytosolic face of the ER membrane, as has been reported for the T-cell receptor ␣-chain and other ER luminal and secretory proteins (29 -36). It would then accumulate upstream in the secretory pathway in the ER and Golgi and also diffuse to the nucleus from the cytosol when the degradation is blocked. Recently, cytosolic accumulation of two transmembrane proteins, presenilin-1 and cystic fibrosis transmembrane regulator, has been described upon inhibition of proteasomal function (37). We do not observe significant accumulation of PrP 145 in the cytosol after proteasomal inhibition. Instead, PrP 145 seems to be specifically targeted to the nucleus by a nuclear localization signal that becomes functional when the carboxyl end of the protein is truncated at residue 145. One type of nuclear localization sequence comprises one or more clusters of basic amino acid residues, which, however, lack tight consensus sequence (38). Interestingly, the N terminus of PrP has a cluster of amino acids (KKRPKP) similar to the SV-40 large T antigen nuclear localization signal (PKKKRKV). Studies are ongoing to establish if this sequence functions as a cryptic nuclear localization signal.
We did not detect ubiquitinated PrP 145 even though our data prove conclusively that the proteasomal pathway degrades PrP 145 . Whether PrP 145 is degraded without ubiquitination as observed for other proteins (29) or is tagged by some other ubiquitin-like protein (39) remains to be determined.
Applicability of the Present Model to the 145 GSS Human Disease-The PrP 145 forms present several unusual characteristics when they are compared with the other mutant PrPs expressed in transfected cells (13,14,19). First, they are highly unstable and are for the most part rapidly degraded. Second, when degradation is impaired, they become partially resistant to protease digestion. Third, paradoxically, the PrP 145 is more protease-resistant in the dispersed than in the aggregated form and, in this form, the protease-resistant fragments include the intact N terminus, whereas a 1.5-2-kDa sequence located at the C terminus remains protease-sensitive. This contrasts with the data from the other transfected cell models of inherited prion diseases in which the mutant PrP spontaneously becomes protease-resistant, the protease-resistant fraction is present only in the aggregated form, and the protease-resistant core includes residues ϳ90 -231 (14,19,41). The present findings argue that also the N-terminal region of PrP, including the signal peptide, may aggregate and become weakly resistant to proteases. Moreover, they indicate that because PrP 145 is also resistant to protease treatment in the dispersed state, the protease resistance is not exclusively because of aggregation but to other mechanisms such as, for example, the presence of a protective ligand or the adoption of a ␤-sheet conformation in a monomeric or oligomeric state. A recent report supports this assumption (47). It has also been recently shown that the PrP 121-231 C terminus segment can adopt a ␤-sheet conformation at acidic pH, independently of other segments (40).
The salient histopathological features of the human Y145stop variant of GSS are the widespread PrP amyloid deposits in vessels and parenchyma of brain and the presence of intraneuronal fibrillary inclusions called neurofibrillary tangles, whereas spongiform degeneration is lacking (7,11,12). The amyloid deposits have been shown to immunostain with antibodies raised to the N-terminal 25 amino acids of PrP, indicating the presence of N-terminal fragment(s) of PrP 145 (11). In addition, an N and C terminus-truncated ϳ7.5-kDa PrP fragment has been detected in monomeric and oligomeric  forms, which by epitope mapping is believed to include amino acids 90 -147 (12). A ϳ7.5-kDa PrP fragment has also been isolated from the amyloid deposits of other GSS variants associated with PRNP point mutations, and it has been found to be the only PK-resistant PrP form recovered from the brain when spongiform degeneration is absent (8,10,42). In the P102L GSS variant, the ϳ7.5-kDa fragment has been shown to span residues 78 -82 to residues 147-150 by sequence and mass spectrophotometric analyses (42). Therefore, the 7.5-kDa fragment present in the amyloid deposits of Y145stop GSS variant is likely to include residues ϳ80 to ϳ145 and to be the only major PK-resistant PrP fragment present in the brain parenchyma of subjects affected by this disease.
It is not immediately evident how the findings of the human disease and the present findings can be reconciled. We did not find a 7.5-kDa PrP fragment or any fragment of smaller size. Data obtained from cell models of inherited prion diseases have been compared with those obtained from brains affected by the corresponding disease in a previous study (13). It was found that although the cell model does not form a PK-resistant PrP comparable with that of the disease, it reproduces the metabolic changes occurring in the mutant PrP in the brain (13). Therefore, it is reasonable to postulate that PrP14 and 15.5 forms are expressed in the brain of the subjects carrying the Y145stop PRNP mutation and are in large amount cleared through the proteasomal pathway. Effective proteasomal degradation of PrP 145 along with the presence of the PrP C encoded by the normal allele may prevent the expression of disease until adult age. However, a decrease in proteasomal function with advanced age or the low but continuous intracellular accumulation and secretion of the aggregated and weakly PK resistant PrP 145 would result in the formation of the highly amyloidogenic ϳ7.5-kDa PrP fragment and formation of amyloid deposits. Future studies of Y145stop GSS variant-affected brains should search for the presence and distribution of the PrP14 and 15.5 forms. It would be important to determine whether PrP 145 is present in aggregated and weakly PK-resistant form and whether some of it is located inside the nucleus. These findings would provide indirect evidence that proteasomal degradation is impaired in the human disease.
Other neurodegenerative diseases have also been shown to involve the proteasome. Recently, it has been shown that the proteasomal system participates in the metabolism of amyloid ␤ peptide, the main component of the amyloid accumulating in Alzheimer's disease (33). The presence of PrP 145 in the nucleus also provides an interesting analogy with a group of inherited neurodegenerative diseases, which include Huntington's chorea and forms of cerebellar ataxia. In each of these diseases, the presence of polyglutamine repeat expansions leads the mutated protein to adopt a ␤-sheet structure and to form insoluble, ubiquitinated aggregates in the nucleus (43)(44)(45)(46), consistent with proteasomal involvement in these diseases as well. Studies aimed at evaluating changes in proteasomal function with advancing age will provide important information regarding the role of this organelle in the pathogenesis of these disorders and potential therapeutic approaches.