Approx. 15% of human prion diseases have a pattern of autosomal dominant inheritance, and are linked to mutations in the gene encoding PrP (prion protein), a GPI (glycosylphosphatidylinositol)-anchored protein whose function is not clear. The cellular mechanisms by which PrP mutations cause disease are also not known. Soon after synthesis in the ER (endoplasmic reticulum), several mutant PrPs misfold and become resistant to phospholipase cleavage of their GPI anchor. The biosynthetic maturation of the misfolded molecules in the ER is delayed and, during transit in the secretory pathway, they form detergent-insoluble and protease-resistant aggregates, suggesting that intracellular PrP aggregation may play a pathogenic role. We have investigated the consequence of deleting residues 114–121 within the hydrophobic core of PrP on the aggregation and cellular localization of two pathogenic mutants that accumulate in the ER and Golgi apparatus. Compared with their full-length counterparts, the deleted molecules formed smaller protease-sensitive aggregates and were more efficiently transported to the cell surface and released by phospholipase cleavage. These results indicate that mutant PrP aggregation and intracellular retention are closely related and depend critically on the integrity of the hydrophobic core. The discovery that Δ114–121 counteracts misfolding and improves the cellular trafficking of mutant PrP provides an unprecedented model for assessing the role of intracellular aggregation in the pathogenesis of prion diseases.
- inherited prion disease
- prion protein (PrP)
- protein aggregation
- protein misfolding
Prion diseases are invariably fatal neurodegenerative disorders of humans and animals that arise because of misfolding of the PrPC [cellular isoform of PrP (prion protein)], a GPI (glycosylphosphatidylinositol)-anchored glycoprotein whose function is not clear . Unique among neurodegenerative disorders of protein conformation, prion diseases can be transmitted through the infectious route . The transmissible agent (prion) is composed primarily of PrPSc (scrapie isoform of PrP), a misfolded/aggregated form of PrP that is capable of seeding conformational conversion of PrPC molecules [3,4]. Prion infections are quite frequent in animals, but rare in humans, in whom most prion diseases occur sporadically or are inherited due to mutations in the gene encoding PrP . The genetic forms account for approx. 15% of all human prion diseases, and include fCJD (familial Creutzfeldt–Jakob disease), GSS (Gerstmann–Sträussler–Scheinker) syndrome and FFI (fatal familial insomnia).
A 72-amino-acid insertion in the N-terminus of PrP is associated with a mixed CJD-GSS phenotype , and a replacement of aspartic acid with asparagine at codon 178 is linked to either fCJD or FFI . Transgenic mice carrying mouse PrP homologues of these mutations (PG14 and D177N respectively) develop fatal neurological illnesses that model key features of the corresponding human disorders [8–10]. These mice synthesize misfolded forms of mutant PrP in their brains, with biochemical properties reminiscent of PrPSc, including insolubility in non-denaturing detergents, resistance to mild PK (proteinase K) digestion and resistance to cleavage of their GPI anchor by phospholipase [8,9]. Analysis of PG14 and D177N PrP metabolism in primary neurons shows that the biosynthetic maturation of these mutants in the ER (endoplasmic reticulum) is delayed  and they accumulate abnormally in the ER and Golgi apparatus [9,12].
These observations suggest ways by which mutant PrP might damage neurons . Excessive accumulation of misfolded PrP in secretory organelles might saturate the protein folding and transport machinery leading to defective secretion of proteins important for normal neuronal function . Additionally, build-up of mutant PrP in the ER lumen could stimulate stress-response mechanisms, such as the unfolded protein response, leading to neuronal death . Alternatively, neuronal dysfunction might result from toxicity of mutant PrP molecules that escape quality-control systems in the secretory pathway and reach the cell surface, where they may interact abnormally with other proteins, such as ion channels, receptors or signalling molecules . With a view to determining whether the toxic PrP molecules are those retained inside the cell or those that escape, we sought ways to influence mutant PrP misfolding and intracellular trafficking.
The central HC (hydrophobic core) region between amino acids 111 and 135 plays a crucial role in PrP misfolding. After conformational transition to the PrPSc state, this region becomes inaccessible to antibodies, suggesting that it undergoes a profound structural rearrangement . Consistent with this view, synthetic PrP peptides that overlap the HC have an intrinsic tendency to spontaneously adopt a β-sheet-rich structure reminiscent of PrPSc [18–20]. Conversely, PrP molecules that are deleted in the HC have greater conformational stability than WT (wild-type) PrP  and are refractory to PrPSc-induced conversion [22,23]. We tested whether deleting part of the HC of PG14 and D177N PrPs antagonized their intrinsic propensity to misfold and aggregate within the cell.
In the present paper, we report that PG14 and D177N PrPs harbouring the 114–121 deletion (hereafter referred to as PG14/ΔHC and D177N/ΔHC) have less tendency to aggregate and accumulate in intracellular compartments, and are delivered to the cell surface more efficiently than their full-length counterparts. This indicates that the HC region is involved in misfolding, oligomerization and intracellular retention of mutant PrP. Comparison of the neurotoxicity of PG14/ΔHC and D177N/ΔHC and their full-length counterparts may shed light on how intracellular accumulation and cell-surface expression of mutant PrP contribute to the pathogenesis, indicating effective modalities for therapeutic intervention.
MATERIAL AND METHODS
cDNAs encoding mouse WT, PG14 and D177N PrPs containing the 3F4 epitope tag have been described previously . Amino acids 114–121 (WT PrP numbering) were deleted using the GeneTailor™ Site-Directed Mutagenesis System (Invitrogen) with partially overlapping primers: 5′-ATGAAGCATATGGCAGGGGGGGGCCTTGGTGGC-3′ (forward) and 5′-CCCTGCCATATGCTTCATGTTGGTTTTTGG-3′ (reverse). The WT and PG14 PrP constructs containing a monomerized version of EGFP (enhanced green fluorescent protein) inserted after codon 34 of mouse PrP have been described previously . To generate PrP–EGFP fusion molecules carrying Δ114–121 alone or with the PG14 mutation, KpnI fragments from the 5′ end of WT PrP–EGFP and PG14 PrP–EGFP constructs were purified and ligated to the 3′ end of PrP Δ114–121 digested with the same enzyme. A cDNA construct encoding 3F4-tagged mouse PrP in which the GPI anchor attachment signal had been replaced with the transmembrane domain of the mannose 6-phosphate receptor (PrP-M6P) was generously provided by Professor David A. Harris (Department of Biochemistry, Boston University School of Medicine, Boston, MA, U.S.A.) and will be described in a forthcoming publication (details available on request from Professor David A. Harris). The identity of all constructs was confirmed by sequencing the entire coding region. All constructs were cloned into pcDNA3.1(+)/Hygro expression plasmid (Invitrogen).
BHK (baby hamster kidney)-21 cells (A.T.C.C.; CCL-10) were grown in MEM (minimal essential medium)-α (Gibco) supplemented with 10% (v/v) FBS (fetal bovine serum), 100 μM non-essential amino acids, 1× minimal essential vitamins, 100 units/ml penicillin and 100 μg/ml streptomycin. HEK (human embryonic kidney)-293 cells (A.T.C.C.; CRL-1573) were grown in a 1:1 mixture of MEM-α and DMEM (Dulbecco’s modified Eagle’s medium) (Gibco) containing 10% (v/v) FBS, 2 mM glutamine, non-essential amino acids and penicillin/streptomycin. Cells were transfected with pcDNA3.1(+)/Hygro expression plasmids using Lipofectamine™ 2000 (Invitrogen). Stable transfected clones were selected with 200 μg/ml hygromycin for 10–14 days, then individual clones were isolated and maintained with 100 μg/ml hygromycin. The results of the present study were obtained from at least two independent cloned lines expressing each construct.
Cerebellar granule neurons were prepared from 6-day-old C57BL/6J mice, as described in . All procedures involving animals and their care were conducted according to European Union (EEC Council Directive 86/609, OJ L 358, 1; 12 December 1987) and Italian (D.L. n.116, G.U. suppl. 40; 18 February 1992) laws and policies, and in accordance with the United States Department of Agriculture Animal Welfare Act and the National Institute of Health (Bethesda, MA, U.S.A.) policy on Humane Care and Use of Laboratory Animals, and they were approved by the Animal Care Committee of the Department of Neuroscience at Mario Negri Institute (Milan, Italy). Cerebella were dissected, sliced into ~1 mm pieces and incubated in HBSS (Hanks balanced salt solution) (Gibco) containing 0.3 mg/ml trypsin (Sigma–Aldrich) at 37 °C for 15 min. Trypsin inhibitor (Sigma–Aldrich) was added to a final concentration of 0.5 mg/ml and the tissue was mechanically dissociated by passing through a flame-polished Pasteur pipette. Cells were transfected using the Nucleofector® device and the Mouse Neuron Nucleofector® Kit (Lonza) following the manufacturer’s instructions. After transfection, cells were pelleted, resuspended in RPMI 1640 (Gibco) containing 10% (v/v) dialysed FBS (Sigma–Aldrich) and 2 mM glutamine, and plated at 450000 cells/cm2 on poly-L-lysine (0.1 mg/ml)-coated plates. After 2 h, the medium was replaced with BME (Basal Medium Eagle) (Gibco) supplemented with 10% (v/v) FBS, penicillin/streptomycin (as above) and 25 mM KCl, and cells were maintained at 37 °C in an atmosphere of 5% CO2/95% air. Cells were analysed after 7–10 days in culture.
Monoclonal antibody 3F4  was used at a 1:2000 dilution for Western and slot blotting, and at a 1:500 dilution for immunofluorescence staining and cell blotting. Monoclonal antibody 12B2  was used at a 1:5000 dilution for Western blotting. Anti-giantin (Covance) and anti-PDI (protein disulfide isomerase) (Sigma–Aldrich) rabbit polyclonal antibodies were used at 1:1000 and 1:500 dilutions respectively for immunofluorescence staining.
Samples were diluted in Laemmli sample buffer (4% SDS, 20% glycarol, 10% 2-mercaptoethanol, 0.004% Bromophenol Blue and 0.125 M Tris/HCl), heated at 95 °C for 5 min then resolved by SDS/PAGE (12% gel). Proteins were electrophoretically transferred on to PVDF membranes and the membranes were blocked for 10 min in 5% (w/v) non-fat dried skimmed milk powder in Tris-buffered saline containing 0.05% Tween 20. After incubation with appropriate primary and HRP (horseradish peroxidase)-conjugated secondary antibodies, signals were revealed using enhanced chemiluminescence (Amersham Biosciences) and visualized using an XRS image scanner (Bio-Rad Laboratories). Anti-PrP antibodies 3F4 or 12B2 were used to develop Western blots, as indicated in the Figure legends. Quantitative densitometry of protein bands was performed using Quantity One software (Bio-Rad Laboratories). A one-tailed unpaired Student's t test was used for statistical analysis.
Detergent insolubility and PK resistance
To assay detergent insolubility, cells were lysed in LB (lysis buffer: 10 mM Tris/HCl, pH 7.5, 100 mM NaCl, 0.5% sodium deoxycholate and 0.5% Nonidet P-40), containing a protease inhibitor cocktail (SigmaFAST™ S8820; Sigma–Aldrich). After a brief centrifugation to remove debris, lysates corresponding to 300 μg of protein were centrifuged at 55000 rev./min for 45 min in a Beckman Optima Max-E ultracentrifuge using a TLA-55 rotor. Proteins in the pellet and supernatant were separated by SDS/PAGE (12% gel) and analysed by Western blotting. To assay PK resistance, cell extracts prepared in LB at a concentration of 1 mg/ml of total proteins were incubated with PK at 4 °C. Compared with the standard incubation at 37 °C, this procedure allows a better detection of weakly protease-resistant forms of mutant PrP [9,26].
Immunoprecipitation with 15B3
A 100-μl sample of mouse anti-IgM antibody–Dynabeads® (Dynal) and 50 μg of monoclonal antibody 15B3 (Prionics) were diluted with 1 ml of PBS containing 0.1% BSA, incubated for 2 h, washed three times with PBS and resuspended in 100 μl of PBS; 10 μl of 15B3-coated Dynabeads® were then added to the cell lysates in LB containing 500 μg of total protein. Samples were incubated on a rotating wheel for 24 h at 4 °C, after which beads were washed twice with 1 ml of Wash Buffer (Prionics). Immunoprecipitated PrP was eluted by boiling in Laemmli sample buffer and analysed by Western blotting.
Sedimentation of PrP in sucrose gradients
Cell lysates were diluted to a final concentration of 0.5 mg/ml (total protein) in LB supplemented with protease inhibitors, incubated for 20 min at 4 °C and centrifuged at 13000 g for 5 min. A 0.5-ml portion of the cleared samples was fractionated on a 5-ml step sucrose gradient (10, 30, 40, 50 and 60%) in LB by ultracentrifugation at 45000 rev./min at 4 °C in an MLS-50 rotor using an Optima MAX-E ultracentrifuge (Beckman). After centrifugation, 0.5-ml fractions of the gradient were collected, and proteins in each fraction were methanol-precipitated and analysed by Western blotting.
Immunofluorescence staining and imaging
Cells were seeded on to glass coverslips and grown for 24 h to 50% confluence. For surface staining of PrP, cells were washed with ice-cold PBS and incubated for 1 h at 4 °C with antibody 3F4 diluted 1:500 in Opti-MEM® (Life Technologies), followed by washing with PBS and fixation in 4% paraformaldehyde. Cells were then washed with PBS, incubated with blocking solution (0.5% BSA and 50 mM NH4Cl in PBS) for 1 h at room temperature (25 °C) with Alexa Fluor® 488-conjugated goat anti-(mouse IgG) antibody (Molecular Probes) diluted 1:500 in blocking solution. For co-localization experiments, cells grown on glass coverslips were washed with PBS and fixed for 30 min at room temperature with 4% paraformaldehyde in PBS. Cells were then washed with PBS, incubated with blocking solution containing 0.1% saponin, and incubated with the primary and Alexa Fluor® 488- or 594-conjugated secondary antibodies diluted in the same solution. In some experiments, cells were stained with 1 μg/ml DAPI (4′,6-diamidino-2-phenylindole) (Sigma–Aldrich) for 5 min. Coverslips were mounted with a gel mount (Biomedia), and analysed with an Olympus FV500 laser confocal scanning system. For imaging PrP–EGFP, cells were grown on μ-Dishes (Ibidi) and viewed with a CellR imaging station (Olympus) coupled to an inverted microscope (IX 81; Olympus). The EGFP fluorescent signal was acquired with a high-resolution camera (ORCA) equipped with a 488 nm excitation filter and an emission filter with a range of 510±40 nm. The DAPI signal was acquired with a 390 nm excitation filter and an emission filter with a range of 430±20 nm.
Assay of PIPLC (phosphatidylinositol-specific phospholipase C) release
For PIPLC-release experiments, cells grown on coverslips were incubated with 4 μg of recombinant PIPLC (Prozyme) for 2 h at 37 °C, followed by surface staining for PrP. PrP released into the medium was quantified by slot blot analysis. After incubation with or without PIPLC, the medium was collected and the adherent cells washed with PBS and lysed in LB. Samples were then transferred on to a nitrocellulose membrane using a slot blot apparatus (both from Bio-Rad Laboratories), and probed with anti-PrP antibody 3F4 and HRP-conjugated anti-(mouse IgG) secondary antibody. The signals were revealed by enhanced chemiluminescence, visualized by a Bio-Rad Laboratories XRS image scanner and quantified using Quantity One software.
The cell blotting technique we used is a modified version of the procedure described by Bosque and Prusiner . Cells grown to confluence on glass coverslips were washed twice with PBS. For detection of surface PrP, coverslips were placed on ice and incubated for 20 min in Opti-MEM® (Life Technologies) with antibody 3F4 (this step was omitted for detection of total PrP). Coverslips were subsequently removed and placed on Whatman 3MM chromatography paper. A nitrocellulose membrane (Bio-Rad Laboratories) was first immersed in distilled water, then soaked in blotting buffer (0.5% sodium deoxycholate, 0.5% Nonidet-P40 and 50 mM Tris/HCl, pH 7.4) and placed on top of the coverslips. The sandwich was backed with blotting-buffer-soaked chromatography paper and pressed firmly for 1 h. The coverslips were carefully removed from the membrane and the blot was blocked for 10 min in 5% (w/v) non-fat dried skimmed milk powder in Tris-buffered saline containing Tween 20. After incubation with the 3F4 antibody (this step was omitted for cells previously labelled with 3F4) and HRP-conjugated secondary antibody, signals were revealed using enhanced chemiluminescence (Amersham Biosciences) and visualized with a Bio-Rad Laboratories XRS image scanner. Quantitative densitometry was performed using Quantity One software.
PG14 PrP molecules carrying Δ114–121 have a reduced propensity to misfold and aggregate
In transfected cells, PG14 PrP misfolds spontaneously and forms detergent-insoluble aggregates that are resistant to mild PK digestion [28–31]. On the basis of observations that deletion of amino acids 114–121 stabilizes PrP conformation , we tested whether Δ114–121 counteracted the tendency for PG14 PrP to misfold. PrP cDNAs carrying Δ114–121 alone (ΔHC) or in combination with the PG14 mutation (PG14/ΔHC) were generated by in vitro mutagenesis and expressed in HEK-293 cells. The PrP level in transfected HEK-293 cells was similar to that of endogenous PrP in cerebellar granule neurons from C57BL/6 mice (Figures 1A and B). The glycosylation patterns of ΔHC and PG14/ΔHC PrPs were indistinguishable from those of WT and PG14 PrPs (Figure 1C). After incubation with N-glycosidase F to remove the N-linked oligosaccharides, ΔHC and PG14/ΔHC PrP had an apparent molecular mass approx. 1 kDa lower than their full-length counterparts because of the eight-amino-acid difference between the polypeptide chains (Figure 1C; compare lanes 5 and 6, and 7 and 8).
To test whether deleting residues 114–121 affected PG14 PrP aggregation, detergent cell extracts were ultracentrifuged, and the amount of PrP in the supernatant and pellet was determined by Western blotting. Consistent with previous analyses [12,30], approx. 40% of PG14 PrP was aggregated, partitioning in the pellet fraction (Figures 2A, lanes 5 and 6, and 2B). Deleting amino acids 114–121 dramatically improved the solubility, with only ~15% of PG14/ΔHC PrP in the insoluble fraction (Figures 2A, lanes 7 and 8, and 2B). Results were similar when the aggregated PrP was measured by 15B3 immunoprecipitation, which selectively detects a variety of aggregated forms of PrP, including PG14 [31,32] (Figures 2C and 2D). Finally, we tested the effect of Δ114–121 on the protease resistance of PG14. Detergent cell extracts were incubated with PK at 4 °C to enhance detection of weakly PK-resistant forms of PrP [9,26]. PG14/ΔHC was PK-sensitive, in contrast with the mild protease resistance of PG14 PrP (Figure 3). There were no differences in solubility, 15B3 reactivity or protease sensitivity between WT and ΔHC PrPs (Figures 2 and 3).
PG14/ΔHC forms smaller aggregates than PG14 PrP
The differences in solubility, 15B3 reactivity and PK resistance between PG14 and PG14/ΔHC suggested that the two molecules might differ in their states of aggregation (quaternary structure). To compare the size distribution of PG14 and PG14/ΔHC PrP aggregates, we fractionated lysates of stably transfected HEK-293 cells by velocity sedimentation on 0–60% sucrose gradients and analysed PrP in the different fractions by Western blotting. WT and ΔHC PrP sedimented entirely in fractions 1–4 of the gradient (Figures 4A, 4B, 4E and 4F). In contrast, PG14 PrP was present in all fractions, with a large amount of molecules in the lower part of the gradient (fractions 7–10) (Figures 4C and 4G). Deletion of residues 114–121 dramatically changed the sedimentation properties of the mutant protein. Similarly to WT PrP, PG14/ΔHC distributed in the top half of the gradient, with the majority of molecules in fractions 1–4 and a small amount in fraction 5 (Figures 4D and 4H). These results indicated that PG14 and PG14/ΔHC differed in their quaternary state, the latter forming much smaller aggregates.
PG14/ΔHC PrP is efficiently expressed on the cell surface and released by phospholipase cleavage of its GPI anchor
PG14 PrP is delayed in its export from the ER and at steady state is present on the cell surface at lower levels than WT PrP [11,12,33]. To investigate the effect of HC deletion on PrP trafficking, we analysed the protein distribution in transiently transfected BHK-21 cells in which the subcellular localization of WT and PG14 PrP had been characterized . To visualize PrP on the plasma membrane, live cells were incubated at 4 °C with anti-PrP antibody 3F4, fixed and reacted with Alexa Fluor® 488 (green)-conjugated secondary antibody without permeabilization. To visualize intracellular PrP, cells were fixed and permeabilized before incubation with 3F4. Permeabilized cells were also reacted with antibodies against PDI or giantin as markers of the ER and Golgi respectively. WT and ΔHC PrPs were efficiently expressed on the cell surface (Figures 5A and 5B) and partly co-localized with PDI and giantin (Figures 5E, 5F, 5I and 5L, yellow colour). PG14 PrP was weakly detected on the plasma membrane (Figure 5C) and had a perinuclear intracellular distribution that, consistent with previous results [12,33], localized with ER and Golgi markers (Figures 5G and 5M). PG14/ΔHC PrP, in contrast, showed cell-surface localization similar to WT and ΔHC PrPs (Figure 5D). However, compared with WT and ΔHC, a higher fraction of PG14/ΔHC was found in the ER and Golgi (Figures 5H and 5N, yellow colour).
Results were similar when PrP distribution was investigated in transfected HEK-293 cells, where we quantified the amount of surface PrP using a cell blotting technique. This confirmed that PG14/ΔHC PrP was expressed on the cell surface at significantly higher levels than PG14, and that there was no difference in surface expression between WT and ΔHC PrPs (Figure 6).
Next, we analysed PrP distribution in neurons using PrP– EGFP fusion molecules. Primary cultures of cerebellar granule neurons from C57BL/6J mice were transfected with plasmids encoding WT PrP–EGFP, PG14 PrP–EGFP or PG14/ΔHC PrP–EGFP. We imaged PrP–EGFP proteins in fixed DAPI-stained cells to analyse their localization with respect to the nucleus. Consistent with previous immunolocalization of non-fluorescent versions of WT and PG14 PrP in neurons , WT PrP–EGFP distributed on the cell soma and along the neurites (Figure 7A), with PG14 PrP–EGFP mainly in intense perinuclear patches (Figure 7B). The cellular distribution of PG14/ΔHC PrP–EGFP was similar to WT PrP–EGFP (Figure 7C), indicating that disrupting the HC domain prevented intracellular aggregation of PG14 PrP, improving its trafficking to the neuronal surface.
WT PrP can be efficiently released from cell membranes by treatment with the bacterial enzyme PIPLC, which cleaves the glycerolipid portion of the GPI anchor. In contrast, PG14 PrP is PIPLC-resistant [8,30,34–36]. To test whether the GPI anchor of PG14/ΔHC PrP was susceptible to cleavage, live HEK-293 cells were incubated with or without PIPLC and the surface was immunostained with anti-PrP antibody. As with WT and ΔHC PrPs, there was a clear fall in surface PG14/ΔHC PrP immunofluorescence after PIPLC treatment, indicating that the GPI anchor was efficiently cleaved and the protein released from the plasma membrane (Figures 8A and 8F, 8B and 8G, and 8D and 8I). Consistent with previous results [34,35], the small amount of PG14 PrP on the cell surface was not released by PIPLC (Figures 8C and 8H). As expected, PrP-M6P, an engineered form of PrP in which the GPI anchor attachment signal had been replaced with the transmembrane domain of the mannose-6-phosphate receptor protein, was PIPLC-resistant (Figures 8E and 8L). Quantification of PrP in the medium confirmed that PIPLC treatment did not release PG14 and PrP-M6P from the cell surface, whereas it efficiently released WT, ΔHC and PG14/ΔHC PrPs (Figure 8M).
Deletion of the HC domain increases solubility and cell-surface expression of D177N PrP
Next, we tested whether deleting the HC region affected the biochemical properties and cellular distribution of another pathogenic PrP mutant. Like PG14 PrP, mouse PrP carrying the D177N substitution has abnormal biochemical properties, altered cellular localization [11,12,14,28,33,36] and is pathogenic in transgenic mice [9,10]. When transiently expressed in HEK-293 cells, ~30% of D177N PrP was detergent-insoluble (Figures 9A, lanes 1–2, and 9B), and selectively immunoprecipitated by the 15B3 antibody (Figures 9C, lanes 1–2, and 9D). In contrast, only a small amount of D177N/ΔHC PrP (~10%) was insoluble (Figures 9A, lanes 3–4, and 9B) and 15B3-reactive (Figures 9C, lanes 3–4, and 9D). Surface immunofluorescence staining of live (unpermeabilized) cells indicated that D177N/ΔHC PrP was expressed on the plasma membrane at a higher level than D177N PrP (Figures 10A and 10B). Immunofluorescence analysis after permeabilization to detect also intracellular PrP showed that, consistent with previous results [12,14], D177N PrP was mostly concentrated in perinuclear patches co-localizing mainly with the Golgi marker giantin (Figures 10C and 10E). In contrast, D177N/ΔHC was present more on the cell surface and showed a more diffuse intracellular distribution, co-localizing with both ER and Golgi markers (Figures 10D and 10F). Thus Δ114–121 counteracted aggregation and improved intracellular trafficking of D177N PrP, as with the PG14 mutation.
NMR analysis of recombinant and brain-derived PrP showed that residues 23–125 form a flexibly disordered tail, and residues 128–230 a globular domain comprising three α-helices and two short β-strands flanking helix 1 [37,38]. The unstructured portion of PrP between residues 90 and 120 undergoes profound rearrangements in PrPSc  and spontaneously forms amyloid when synthesized as a peptide [18–20], indicating that this region is essential to PrP refolding and aggregation. Consistent with this view, PrP molecules deleted in the HC region between residues 109 and 112 or 114 and 121 are refractory to conformational conversion induced by PrPSc and have a trans-dominant inhibitory effect on conversion of WT PrP [22,23].
In the present study, we found that the biochemical properties and cellular trafficking of WT and Δ114–121 PrPs are similar. This, and the evidence that PrP Δ114–121 is not toxic in mice and retains the neuroprotective activity of WT PrP , supports the use of this molecule in gene therapy of prion infections .
We also found that Δ114–121 antagonized the spontaneous conversion of two mutant PrPs carrying either 72 extra amino acids in the N-terminus or a single residue substitution in the C-terminal domain. Thus the HC of PrP is vital for protein misfolding, independently of whether this is triggered by interaction with PrPSc or by structural information intrinsic to mutant PrP. The fact that region 114–121 governs both mutation- and PrPSc-induced PrP misfolding suggests that similar structural alterations underlie genetic and infectious prion diseases, a conclusion also emerging from other studies . Since PrPSc-induced conversion is likely to occur in the endosomal recycling compartment , whereas mutant PrP first adopts an abnormal conformation in the ER , our data also suggest that region 114–121 influences PrP misfolding independently of the intracellular milieu.
In transfected cells, PG14 PrP conversion into the PrPSc-like state proceeds stepwise through a series of identifiable biochemical intermediates . The initial alteration in the mutant molecule occurs in the ER, during or immediately after translation of the polypeptide chain, and is manifested by acquisition of PIPLC resistance. This is likely to reflect PG14 PrP misfolding and oligomerization, since the GPI anchor becomes sensitive to cleavage after protein denaturation . PG14 PrP aggregates further in post-Golgi locations where it first becomes detergent-insoluble, then PK-resistant, very probably because of an incremental change in aggregation [29,43]. We found that most PG14/ΔHC PrP molecules expressed in HEK-293 cells were susceptible to PIPLC cleavage, suggesting that Δ114–121 inhibited the initial step of conformational conversion in the ER, probably by introducing an energy barrier to misfolding . Δ114–121 also had a dramatic effect on aggregation. Cells expressing PG14/ΔHC and D177N/ΔHC PrPs accumulated only small amounts of detergent-insoluble and 15B3-reactive molecules, which sedimented in the lighter fractions of sucrose gradients and were PK-sensitive. These results indicate that the HC deletion changed the aggregation properties of mutant PrP, favouring the formation of small, weakly associated, oligomers. There is evidence that intermolecular PrP interactions involved in forming β-pleated sheet are dependent on specific residues in the central hydrophobic region [44,45]. These residues are lacking in HC-deleted molecules, suggesting that the intermolecular interactions necessary for formation and/or stabilization of the aggregates may be disrupted.
Several different pathogenic mutations share a common effect on PrP trafficking, impairing delivery to the cell surface and causing a portion to accumulate in intracellular compartments [9,11,12,14,33,46–53]. Δ114–121 greatly improved the secretory trafficking of PG14 and D177N PrPs. Immunofluorescence microscopy in BHK-21 and HEK-293 cells and imaging of PrP–EGFP fusion proteins in live neurons showed that PG14/ΔHC and D177N/ΔHC PrPs were efficiently expressed on the cell surface. This observation, in conjunction with the biochemical results, indicates that the effect of the HC deletion on PrP structure and cellular trafficking are closely related, and that impaired trafficking may be a direct consequence of PrP misfolding and aggregation. However, it is possible that Δ114–121 prevents intracellular accumulation through a different mechanism, for example by disrupting a structural motif of PrP essential for recognition and retention of misfolded molecules in the secretory pathway. In this regard, it could be interesting to study the effect of Δ114–121 on the trafficking of PrP mutants that accumulate in the secretory pathway, but are less prone to aggregation and less PK-resistant, such as E200K and P102L [28,33,49].
The cellular pathways activated by mutant PrP ultimately leading to neuronal dysfunction and degeneration in genetic prion diseases are poorly understood. In different non-neuronal and neuronal cells, several mutant PrPs misfold in the early steps of the secretory pathway, reside longer in the ER and Golgi apparatus, and are delivered less efficiently to the cell surface, suggesting that intracellular PrP accumulation may be pathogenic. The discovery that Δ114–121 counteracts misfolding and improves the cellular trafficking of mutant PrP provides an unprecedented model for assessing the role of intracellular retention in neurotoxicity. By comparing the effects of HC-deleted and full-length PG14 and D177N molecules on neuronal viability, it should be possible to assess the contributions of intracellular accumulation and cell-surface localization. Expression of PG14 or D177N PrP in cultured cells does not significantly alter cell viability, growth rate or susceptibility to toxic drugs (, and T. Massignan, E. Restelli, E. Biasini and R. Chiesa, unpublished work); thus the only way to test the effect of Δ114–121 on toxicity is to study the consequences of PG14/ΔHC or D177N/ΔHC expression in transgenic mice. Since Δ114–121 alone does not induce neurodegenerative illness in mice , any deviation from the phenotype of mice expressing the full-length mutants might be ascribable to changes in PrP aggregation and/or trafficking.
It is interesting to speculate on the possible outcome of this experiment. Should PG14/ΔHC and D177N/ΔHC turn out not to be toxic, this would indicate that intracellular aggregation is vital in the disease process, and that therapeutic strategies aimed at favouring mutant PrP folding and trafficking could be beneficial [55–57]. Should the deleted molecules be more toxic than their full-length counterparts, this would imply that mutant PrP expression on the plasma membrane and/or formation of small oligomers are primarily responsible for toxicity.
In conclusion, the present study indicates that sequence 114–121 within the unstructured region of PrP is vital for protein misfolding and aggregation, and sets the basis for future investigations to ascertain the role of intracellular accumulation of misfolded PrP in the pathogenesis of inherited prion diseases.
Emiliano Biasini, Laura Tapella, Elena Restelli, Manuela Pozzoli and Tania Massignan performed the experiments. Emiliano Biasini and Roberto Chiesa conceived and designed the experiments, analysed the data and wrote the paper.
This work was supported by the Fondazione Telethon [grant number TCR08005 (to R.C.)]; and by a fellowship from the Fondazione Monzino to E.B. R.C. is an Associate Telethon Scientist (Dulbecco Telethon Institute, Fondazione Telethon).
We thank Richard Kascsak (New York State Institute for Basic Research, Staten Island, NY, U.S.A.) for the 3F4 antibody, Jan P. Langeveld [Central Veterinary Institute (CVI), Wageningen University and Research Centre, 8200 AB Lelystad, The Netherlands] for the 12B2 antibody, Alex Raeber and Bruno Oesch (Prionics, Zürich, Switzerland) for the 15B3 antibody, and David A. Harris for the PrP-M6P cDNA construct. We are grateful to Pietro Veglianese for advice on live cell imaging, to Gianlugi Forloni for discussion and to Judith Baggott for help in preparing the manuscript.
Abbreviations: BHK, baby hamster kidney; DAPI, 4′,6-diamidino-2-phenylindole; EGFP, enhanced green fluorescent protein; ER, endoplasmic reticulum; FBS, fetal borine serum; fCJD, familial Creutzfeldt–Jakob disease; FFI, fatal familial insomnia; GPI, glycosylphosphatidylinositol; GSS, Gerstmann–Sträussler–Scheinker; HC, hydrophobic core; HEK, human embryonic kidney; HRP, horseradish peroxidase; LB, lysis buffer; MEM, minimal essential medium; PDI, protein disulfide isomerase; PIPLC, phosphatidylinositol-specific phospholipase C; PK, proteinase K; PrP, prion protein; PrPC, cellular isoform of PrP; PrP-M6P, PrP with the transmembrane domain of the mannose 6-phosphate receptor; PrPSc, scrapie isoform of PrP; WT, wild-type
- © The Authors Journal compilation © 2010 Biochemical Society