Morikazu Imamura, Prion Disease Research Center, National Institute of Animal Health, 3-1-5 Kannondai, Tsukuba, Ibaraki, 305-0856 Japan. Tel: +81 298 38 7757; fax: +81 298 38 7757; email: email@example.com
The pathogenic isoform (PrPSc) of the host-encoded normal cellular prion protein (PrPC) is believed to be the infectious agent of transmissible spongiform encephalopathies. Spontaneous conversion of α-helix-rich recombinant PrP into the PrPSc-like β-sheet-rich form or aggregation of cytosolic PrP has been found to be accelerated under reducing conditions. However, the effect of reducing conditions on PrPSc-mediated conversion of PrPC into PrPSc has remained unknown. In this study, the effect of reducing conditions on the binding of bacterial recombinant mouse PrP (MoPrP) with PrPSc and the conversion of MoPrP into proteinase K-resistant PrP (PrPres) using a cell-free conversion assay was investigated. High concentrations of dithiothreitol did not inhibit either the binding or conversion reactions of PrPSc from five prion strains. Indeed, dithiothreitol significantly accelerated mouse-adapted BSE-seeded conversion. These data suggest that conversion of PrPSc derived from a subset of prion strains is accelerated under reducing conditions, as has previously been shown for spontaneous conversion. Furthermore, the five prion strains used could be classified into three groups according to their efficiency at binding and conversion of MoPrP and cysteine-less mutants under both reducing and nonreducing conditions. The resulting classification is similar to that derived from biological and biochemical strain-specific features.
10 mM Tris-HCl, pH 7.4 containing 1% sarkosyl, 10% NaCl
Transmissible spongiform encephalopathies are infectious and fatal neurodegenerative diseases of humans and other animals. The conversion of normal host-encoded, PK-sensitive, prion protein (PrPC) into the partially PK-resistant PrPSc pathological form represents the central event in TSE pathogenesis (1). Direct interaction between PrPC and PrPSc is crucial for formation of additional PrPSc from PrPC (2, 3). However, the molecular mechanisms involved remain poorly understood.
Reducing agents facilitate the conformational transition from the α-helix-rich bacterial recombinant PrP into the β-sheet-rich form (4, 5), and a reducing environment in the cytosol promotes either fibrillization or a PrPSc-like conformation of PrPC (3, 6, 7). Thus, reducing conditions likely induce spontaneous conversion of PrPC into either PrPSc or a PrPSc-like form. Alternatively, a free-thiol group may be necessary for PrPSc-dependent conversion in PMCA (8). However, addition of reducing agents inhibited PrPSc-dependent conversion of PrPC into PrPSc-like, PK-resistant PrP (PrPres) in a cell-free conversion assay (9). Thus, the effect of reducing conditions on PrPSc-dependent conversion of PrPC has remained unclear.
To investigate this issue, binding and cell-free conversion assays were performed using MoPrP as a PrPC source and five mouse-adapted prion strain PrPSc as the seed. DTT at concentrations great enough to allow reduction of the disulfide bond did not inhibit binding of MoPrP to PrPSc or conversion of MoPrP into PrPres. Indeed, mBSE-seeded conversion was significantly enhanced. These data suggest that an intracellular reducing environment might accelerate both PrPSc-dependent and spontaneous conversion of PrPC. In addition, the five prion strains were classified according to their efficiency at binding and conversion of MoPrP and the Cys-less mutant in the presence and absence of DTT. This classification correlated well with that based on the pathological and biochemical properties of each strain.
MATERIALS AND METHODS
Transmissible spongiform encephalopathy materials
Mouse scrapie strains Chandler, 79A, ME7, and Obihiro (10) and a mBSE were used. These prion strains were propagated in ICR mice.
An equal volume of 2 × SDS sample buffer was added and samples were boiled for 5 min, followed by resolution by SDS-PAGE using NuPAGE 12% Bis-Tris gels (Invitrogen, Carlsbad, CA, USA) and transferred onto polyvinylidene fluoride membranes. 3F4 antibody (Chemicon, Temecula, CA, USA) and anti-PrP horseradish peroxidase conjugated monoclonal antibody T2 (11) were used for detecting recombinant PrP containing the 3F4 epitope and PK-digested mouse brain-derived PrPSc, respectively. Blotted membranes were developed with SuperSignal West Dura Extended Duration Substrate (Pierce, Rockford, IL, USA), and chemiluminescence signals were detected using a ChemiImager (Alpha InnoTech, San Leandro, CA, USA).
Expression and purification of recombinant prion protein
Full-length mature mouse PrP carrying the 3F4 epitope (amino acids 23–230; MoPrP) was generated by PCR-based site-directed mutagenesis. All amplification reactions were performed using standard PCR conditions. The 5′ portion of MoPrP was amplified from mouse brain-derived cDNA using the following primers: 5′-CATATGAAAAAGCGGCCAAAGCCTG-3′ (5′ forward primer) and 5′-GCCATATGCTTCATGTTGGTTTTTGGTTTG-3′ for a reverse primer containing the 3F4 epitope. The 3′ portion of MoPrP was amplified using the following primers: 5′-AACCAACATGAAGCACATGGCAGGGG-3′ for a forward primer containing the 3F4 epitope and 5′-GGATCCTCATCAGGATCTTCTCCCGTCGTAATAG-3′ for a reverse primer covering the 3′ terminus of MoPrP (3′ reverse primer). The two purified amplicons were combined and used as the template in another PCR with the 5′ forward and 3′ reverse primers. Amplicons were then purified and cloned into a pGEM-T Easy Vector (Promega, Madison, WI, USA).
Two Cys-to-Ser substitution mutants (C213S and C178,213S) were generated by PCR-based site-directed mutagenesis. The primer sets were as follows: for substitution of Cys at 213, 5′-GTACTGGGTGACGCTCATCTGCTC-3′ and 5′-GAGCAGATGAGCGTCACCCAGTAC-3′, and for substitution of Cys at 178, 5′-GTGATATTGACGCTGTCGTGCACG-3′, and 5′-TTCGTGCACGACAGCGTCAATATCAC-3′. PCR to amplify the 5′ and 3′ portions of mutants was performed using the 5′ forward and 3′ reverse primers in combination with the primers above and plasmid cloning MoPrP as a template.
MoPrP, C213S, and C178, 213S were re-cloned from the pGEM-T Easy Vector into pET15b (Novagen, Madison, WI, USA) at NdeI and BamHI sites, and the vectors carrying PrP were transformed into E. coli BL21 (DE3) (Novagen). Expression was carried out according to the manufacturer's instructions. After solubilization of inclusion bodies in binding buffer (0.5M NaCl, 20 mM imidazole, 8 M urea in 20 mM phosphate buffer, pH 7.4), recombinant PrPs were purified under denaturing conditions using a HisTrap HP Kit (Amersham, Arlington Heights, IL, USA) according to the manufacturer's instructions. Purified recombinant PrPs were then dialyzed against 2 M Gdn-HCl and 1 mM EDTA in 50 mM Tris-HCl (pH 8.0). The purities of each PrP were estimated to be >90% by SDS-PAGE and CBB staining. Recombinant PrPs were analyzed by Western blotting with the 3F4 antibody to distinguish recombinant PrP from PrPSc used as seed, and signal intensities were evaluated using a Chemi imager.
Preparation of seed scrapie isoform of prion protein
The scrapie isoform of prion protein was prepared from brain tissue collected from affected animals as described previously (11). Prion-infected mouse brains were homogenized in 10% sarkosyl in 10 mM Tris-HCl (pH 7.4) and then centrifuged at 22,000 g for 10 min. The supernatant was then decanted and centrifuged at 540,000 g for 30 min. The pellets were suspended in TSN with the aid of brief sonication and centrifuged again under the same conditions. The pellets suspended in TSN were treated with 50 μg/mL of PK at 37°C for 60 min. The pellets obtained by centrifugation at 22,000 g for 10 min were washed twice with TSN by centrifugation under the same conditions. The purity of the seed PrPSc fraction was examined by SDS-PAGE and silver staining (Wako, Osaka, Japan). All prion strain PrPSc fractions were adjusted to 200 μg/mL by comparing their signal intensities after Western blotting with that of MoPrP.
Ten micrograms of MoPrP or C213S, and 5 μg of PrPSc derived from the Chandler strain, were incubated in reaction buffer containing DTT or 2ME at 37°C for 24 hr. After incubation, all PrPs were methanol-precipitated and dissolved in 6 M urea in 50 mM Tris-HCl (pH8.0). mBBr was added to a final concentration of 4 mM, and the solutions incubated for 20 min at 25°C to label sulfhydryl groups. Labeled solutions were methanol-precipitated to remove excess mBBr and the pellets dissolved in reducing agent-free SDS-PAGE sample buffer. PrPs were separated by SDS-PAGE, and the gels immersed in 20% methanol/5% acetic acid. The gels were then observed under UV light (365 nm) and stained with Coomassie brilliant blue.
Cell-free conversion reaction assay
A cell-free conversion reaction assay was carried out as described previously (12–14). Seed PrPScs (600 ng) were pretreated with 3 M Gdn-HCl at 37°C for 1 hr. Recombinant PrPs (10 pmol) were added to conversion buffer (1.5 M Gdn-HCl, 1.25% sarkosyl, 50 mM citrate, pH 6.0). Pretreated seed PrPSc was then added to a final reaction volume of 24 μL, and the mixture incubated at 37°C for 24 hr. After incubation, PK (20 μg/mL) digestion was carried out at 37°C for 30 min. The digestion was stopped by adding pefabloc (Roche, Indianapolis, IN, USA) to a final concentration of 2 mM. Proteins in the reaction mixture were recovered by precipitation with four volumes of methanol, and the pellets were dried and dissolved in SDS-PAGE sample buffer containing 100 mM DTT. The recovered de novo PrPress were analyzed by Western blotting using 3F4 mAb, and the signal intensities evaluated using a Chemi imager (Alpha Innotech) . The conversion efficiencies for each sample were expressed as a percentage change (mean ± SE relative to a control value [control = 100]). The Chandler-seeded conversion value of MoPrP under nonreducing conditions was used as a control.
A binding assay was performed as described previously (14). After the same incubation as in the cell-free conversion assay, the reaction mixture was centrifuged at 20,000 g for 10 min at 25°C. The supernatant (unbound fraction) was transferred to a tube, and the pellet (bound fraction) was resuspended in 24 μL of the reaction solution, followed by washing once with 100 μL of the reaction solution. All samples were methanol-precipitated, and SDS-PAGE and Western blotting were conducted as described above. Binding efficiencies were expressed as a percentage change (mean ± SE) relative to control values. The binding value of MoPrP with Chandler PrPSc under nonreducing conditions was used as a control.
The right hemisphere of an end stage prion-affected mouse was fixed in 10% buffered formalin solution. Coronal slices of the brain were cut and immersed in 98% formic acid to reduce infectivity, followed by embedding in paraffin wax. Sections (4 μm) were cut and stained with HE or analyzed by immunohistochemistry. For neuropathological analysis, the lesion profile was determined from the HE-stained sections by scoring the vacuolar changes in nine standard gray-matter areas, as described previously (15). For immunohistochemistry, PrPSc in brain sections was retrieved by the alkaline-based chemical method (16) and then detected using anti-PrP monoclonal antibody SAF84 against the epitope of amino acids 160–170 of the hamster prion protein. Immunoreactions were developed using anti-mouse universal immuno-peroxidase polymer (Nichirei Histofine Simple Stain MAX-PO [M]; Nichirei, Tokyo, Japan) as the secondary antibody and 3′-3 diaminobenzidine tetrachloride as the chromogen.
Reduction of recombinant prion proteins and abnormal prion proteins used as seeds
We investigated whether the disulfide bonds in recombinant wild-type MoPrP and PrPSc are cleaved in the presence of reducing agents. Recombinant PrP and PrPSc were labeled with mBBr following reduction with DTT or 2ME. The fluorescence intensities of mBBr-labeled MoPrP increased in proportion to the reagent concentration; that of MoPrP treated with 100 mM DTT appeared to reach a plateau (Fig. 1a). When the fluorescence signal of 100 mM DTT-treated MoPrP was compared with that of a 100 mM DTT-treated single-Cys substitution mutant (C213S), the signal intensity of the treated MoPrP was about 1.8 times that of treated C213S. We estimated that more than 70% of C213S formed dimers through an intermolecular disulfide bond under nonreducing conditions, but almost all C213S molecules were present as monomers in the presence of 100 mM DTT, suggesting that all C213S molecules had been reduced. As MoPrP contains two Cys residues, its mBBr signal intensity was expected to be twice that of C213S. Therefore, MoPrP was likely reduced almost completely in the presence of 100 mM DTT. Next, we investigated whether Chandler PrPSc was also reduced in the presence of 100 mM DTT (Fig. 1b). Chandler PrPSc was indeed reduced, but only by about 30% (data not shown).
Effects of reducing agents on binding of bacterial recombinant mouse prion protein with scrapie isoform of prion protein and conversion into proteinase K-resistant prion protein
To investigate the effect of reducing conditions on the binding of MoPrP with PrPSc and conversion of MoPrP into PrPres, binding and cell-free conversion assays were first performed using Chandler PrPSc as seed. Addition of both DTT and 2ME resulted in a decrease in the binding and conversion efficiencies in a concentration-dependent manner, but the differences between the reduced and nonreduced samples were not significant (Fig. 2). Addition of another reducing agent, tris(2-carboxyethyl)phosphine, gave similar results (data not shown). These data suggest that reducing conditions did not significantly affect the binding of MoPrP to Chandler PrPSc or conversion of MoPrP into PrPres.
We then investigated the effects of DTT on binding and conversion in several mouse-adapted prion strains. The binding efficiencies of MoPrP with 79A, ME7, Obihiro, and mBSE PrPSc under nonreducing conditions were 104%, 56%, 45%, and 87%, respectively, of that of Chandler (100%) (Fig. 3a, open columns). The efficiencies of ME7 and Obihiro were about half that of Chandler, although there was no significant difference between the two strains and Chandler. On the other hand, the efficiencies of conversion of MoPrP in the 79A, ME7, Obihiro, and mBSE-seeded strains under nonreducing conditions were 94%, 23%, 13%, and 21%, respectively, of that of Chandler. Except for 79A, the differences between Chandler and the other prion strains were significant (P < 0.001) (Fig. 3b, open columns).
The binding and conversion efficiencies of each strain under reducing conditions were compared to those under nonreducing conditions. The binding efficiencies of MoPrP under reducing conditions were not significantly different from those under nonreducing conditions in the four prion strains and Chandler (Fig. 3a, open and solid columns). The efficiencies of conversion of 79A, ME7, and Obihiro and Chandler under reducing conditions were not significantly different from those under nonreducing conditions (Fig. 3b, open and solid columns). However, the efficiency of the mBSE strain was about three times greater under reducing conditions (P < 0.01).
To determine whether the disulfide bond or thiol groups of MoPrP were involved in binding with PrPSc and conversion into PrPres, binding and conversion assays using Cys-less mutants (C178, 213S) were performed. No significant differences were observed in binding efficiencies under nonreducing conditions among C178, 213S, and MoPrP, although the binding efficiencies of C178, 213S with ME7, and Obihiro PrPSc were about half those of MoPrP (Fig. 3a, c, open columns). The efficiency of mBSE-seeded conversion of C178, 213S was not different from that of MoPrP, but the efficiencies of the other strains were lower than that of MoPrP. Furthermore, the difference in the efficiencies of Chandler and 79A were significant (Fig. 3b, d, open columns). Thus, the findings suggested that the effects of Cys to Ser substitutions on binding and conversion were different for each prion strain, and that the presence of Cys or thiol groups was especially important for conversion into PrPres in Chandler and 79A strains.
The binding efficiencies of C178, 213S in each prion strain under reducing conditions were not significantly different from those under nonreducing conditions (Fig. 3c, open and solid columns). The conversion efficiencies of C178, 213S in the Chandler, 79A, ME7, and Obihiro strains under reducing conditions were not significantly different from those under nonreducing conditions, although a significant increase in the conversion efficiency of mBSE was observed under reducing conditions (Fig. 3d, open and solid columns). Therefore, the effect of reducing conditions on the binding and conversion of Cys-less mutants was similar to those of MoPrP, suggesting that neither Cys residues nor thiol groups are involved in the acceleration of mBSE-seeded conversion under reducing conditions.
The biological and biochemical properties of prion strains used as seed
Immunohistochemical and HE staining of brain tissue from mice infected with each strain were performed (Fig. 4a). In mice inoculated with Chandler and 79A strains, diffuse synaptic deposits were found throughout the brain, and the PrPSc accumulation patterns of both strains were very similar. In contrast, ME7 and Obihiro strains produced PrPSc accumulation throughout the neuropil in most areas of the brain, although some areas were predictably severely affected. Large numbers of PrP-aggregates were also detected, but these tended to be small and had less obvious amyloid cores. Furthermore, severe neuronal loss in the CA3 region of the hippocampus and strong vacuolation in the cerebral cortex were observed in these two strains. In the brains of mBSE-inoculated mice, coarse particulate and coalescing types of immunostaining were recognized in the hippocampus and brainstem habenular nuclei. In the cerebral cortex, characteristic lamellar accumulation of PrPSc was detected. In addition, plaque-like deposits were frequently present in the thalamus, corpus callosum, periventricular area, and brain stem of mBSE-inoculated mice. Therefore, the pathological features of each strain group (Chandler and 79A, ME7 and Obihiro, mBSE) were easily distinguishable.
Mean survival times (days ± SD) of mice inoculated with 10% Chandler and 79A, ME7, Obihiro, and mBSE-infected brain homogenates were 141 ± 4.6 and 138 ± 6.9, 150 ± 4.6, 147 ± 2.7, and 160 ± 3.5 days, respectively. Although no significant differences were observed between Chandler and 79A or between ME7 and Obihiro, significant differences in survival times (P < 0.001) were found among the three strain groups.
mBSE and the four scrapie strains, Chandler, 79A, ME7, and Obihiro, could be easily distinguished by their glycoform ratios (Fig. 4b) because the mBSE PrPSc bands migrated faster than scrapie strains. In both the Chandler and 79A strains, monoglycosylated PrPSc predominated, whereas the ME7 and Obihiro strains showed comparable amounts of di- and monoglycosylated protein.
These data suggest that classification of the five strains by biological and biochemical characteristics correlates with that derived from the binding and conversion reactions of each strain.
In this study, we demonstrated that the addition of reducing agents did not inhibit binding and conversion of MoPrP or cysteine-less mutant PrP, and significantly accelerated conversion driven by mBSE PrPSc. Thus, reducing conditions result in an acceleration of PrPSc-dependent conversion in at least some prion strains, as has previously been shown for spontaneous conversion (3–7). Hermann and Caughey reported a contradictory result; they found that addition of DTT decreased conversion by about 90% (9). This may have been due to use of a different recombinant expression system, the origin of the recombinant PrP used as a PrPC source, the prion strains used as PrPSc seed, the preparation method of seed PrPSc, and/or the reaction composition.
Acidic conditions and addition of detergents or denaturants efficiently induce spontaneous conversion of α-helix-rich PrPC into PrPSc-like β-sheet-rich PrP (17, 18). Reducing conditions also stimulate conversion of α-helix-rich recombinant PrP into the β-sheet-rich form (3). In our study, denaturing and mildly acidic reducing reaction conditions were used for the binding and cell-free conversion assays. The conditions in the environment within endosomes and lysosomes, thought to be the location of conversion of PrPC into PrPSc (19–22), are believed to be similar. Addition of DTT significantly enhanced the efficiency of mBSE-seeded conversion, even though DTT did not have a major impact on the conversion efficiency of the other strains. Such differences may be one of the causes of cell tropism for PrPSc accumulation, and furthermore, might result in the prion strain-specific PrPSc accumulation pattern in the brain.
Species specificity in cell-free conversion has been reported (14, 23), and the products preserve strain-specific properties (24). These data suggest that the cell-free conversion reaction mimics some aspects of in vivo conversion of PrPC into PrPSc. In this study, we demonstrated that the effect of reducing conditions and removal of Cys residues on binding and conversion differed among prion strains; indeed, these may mirror prion strain properties in vivo. In fact, classification of the five prion strains by their binding and the conversion efficiencies correlated well with classification according to their biological and biochemical properties. Therefore, the in vivo properties of each strain likely correlate with their conversion capacity. Binding and conversion assays may thus aid in the classification of prion strains.
Reduction of the intramolecular disulfide bond did not interfere with binding of PrPSc to MoPrP and conversion of MoPrP into PrPres. However, substitution of Cys with Ser in MoPrP inhibited binding and conversion of the ME7 and Obihiro strains and conversion of the Chandler and 79A strains. Therefore, Cys residues may play a key role in the conversion and binding of Chandler and 79A, ME7, and Obihiro PrPSc. However, we cannot rule out the possibility that such a substitution alters the tertiary structure of the prion protein.
Addition of DTT significantly increased the conversion efficiencies of MoPrP and the Cys-less mutant driven by mBSE PrPSc. This suggests that the effect of DTT may be mediated by a mechanism other than cleaving of the disulfide bond in MoPrP. DTT diminishes the carbohydrate binding activity of a Cys-less mutant of pigpen as well as inhibiting the intact molecule (25). Therefore, in an mBSE-seeded cell-free conversion, DTT may improve the efficiency of mBSE-seeded conversion independently of the reduction of disulfide bonds.
In summary, reducing conditions did not inhibit conversion in vitro and markedly increased mBSE-seeded conversion. This suggests that cell-free conversion under reducing conditions mimics the conversion of PrPC into PrPSc within endosomes and lysosomes. In addition, classification of prion strains by their efficiency at binding and conversion of both MoPrP and its Cys-less mutant in the absence and presence of DTT correlates well with classification based on biological and biochemical properties. Therefore, the cell-free conversion assay may be useful in discriminating between prion strains.
We are grateful to Dr. Motonori Horiuchi, Graduate School of Veterinary Medicine, Hokkaido University, for donating monoclonal antibody 31C6, and Tomoko Murata and Naomi Furuya for their technical assistance. We also thank the contributions of the animal caretakers. This study was supported by a Grant-in-Aid from BSE Control Project of the Ministry of Agriculture, Forestry and Fisheries of Japan.