Glycosylphosphatidylinositol Anchor-Dependent Stimulation Pathway Required for Generation of Baculovirus-Derived Recombinant Scrapie Prion Protein

Abstract

The pathogenic isoform (PrP^Sc) of the host-encoded cellular prion protein (PrP^C) is considered to be an infectious agent of transmissible spongiform encephalopathy (TSE). The detailed mechanism by which the PrP^Sc seed catalyzes the structural conversion of endogenous PrP^C into nascent PrP^Sc in vivo still remains unclear. Recent studies reveal that bacterially derived recombinant PrP (recPrP) can be used as a substrate for the in vitro generation of protease-resistant recPrP (recPrP^res) by protein-misfolding cyclic amplification (PMCA). These findings imply that PrP modifications with a glycosylphosphatidylinositol (GPI) anchor and asparagine (N)-linked glycosylation are not necessary for the amplification and generation of recPrP^Sc by PMCA. However, the biological properties of PrP^Sc obtained by in vivo transmission of recPrP^res are unique or different from those of PrP^Sc used as the seed, indicating that the mechanisms mediated by these posttranslational modifications possibly participate in reproductive propagation of PrP^Sc. In the present study, using baculovirus-derived recombinant PrP (Bac-PrP), we demonstrated that Bac-PrP is useful as a PrP^C substrate for amplification of the mouse scrapie prion strain Chandler, and PrP^Sc that accumulated in mice inoculated with Bac-PrP^res had biochemical and pathological properties very similar to those of the PrP^Sc seed. Since Bac-PrP modified with a GPI anchor and brain homogenate of Prnp knockout mice were both required to generate Bac-PrP^res, the interaction of GPI-anchored PrP with factors in brain homogenates is essential for reproductive propagation of PrP^Sc. Therefore, the Bac-PMCA technique appears to be extremely beneficial for the comprehensive understanding of the GPI anchor-mediated stimulation pathway.

Transmissible spongiform encephalopathies (TSEs), including scrapie in sheep, bovine spongiform encephalopathy (BSE) in cattle, and Creutzfeldt-Jakob disease in humans, are infectious and fatal neurodegenerative diseases (7). Unique proteinaceous infectious agents called prions are thought to be the cause of TSEs. Prions consist primarily of a pathogenic form (PrP^Sc) of the normal cellular prion protein (PrP^C). PrP^Sc appears to propagate itself by autocatalyzing the conformational conversion of PrP^C (29). Recent findings reveal that PrP^Sc can be amplified in vitro using a protein-misfolding cyclic amplification (PMCA) technique with the bacterial recombinant prion protein (recPrP). The amplified products are infectious and are able to cause TSE in wild-type animals (19, 35), thus providing evidence to support the prion hypothesis. However, the precise mechanism of how endogenous PrP^C converts to nascent PrP^Sc in vivo has yet to be elucidated.

PMCA has been shown to be very effective at detecting minute amounts of scrapie PrP^Sc (31) and is capable of detecting scrapie PrP^Sc in the blood or urine (5, 13, 24, 33). Using brain homogenate (BH) as a PrP^C substrate, PrP^Sc generated by PMCA showed biochemical and biological properties similar to those of the PrP^Sc seed (4). However, the biological properties of bacterial recPrP^Sc generated by PMCA were different from those of the PrP^Sc seed, despite the amino acid sequence of recPrP being identical to that of brain-derived PrP (brain-PrP) (19). Because asparagine (N)-linked glycosylation of PrP^C affects strain-dependent neurotropism significantly (21, 28) and glycosylphosphatidylinositol (GPI) anchoring of PrP^C is necessary for structural conversion and persistent infection in cultured cells (23), these posttranslational modifications should play a central role in the reproductive propagation of PrP^Sc from PrP^C in vivo.

Since N-linked glycosylated (14, 16, 34) and GPI-anchored (9, 25, 30) recombinant proteins can be produced in a baculovirus-insect cell expression system, baculovirus-derived recombinant PrP (Bac-PrP) may work efficiently as the PrP^C substrate, and thus Bac-PrP could be used in vitro as a model of PrP^Sc propagation. Development of an in vitro amplification system using Bac-PrP is of great use for elucidating the molecular mechanisms responsible for the conversion of PrP^C to PrP^Sc mediated by posttranslational modifications. Using the PMCA technique, we report here that both GPI anchoring of Bac-PrP and the factors present in BH in Prnp knockout (Prnp^0/0) mice are required for the generation of PrP^Sc, which preserves the peculiar biochemical and pathological properties of the PrP^Sc seed. The Bac-PrP PMCA system would therefore be a powerful tool for elucidating the GPI anchor-mediated stimulation pathway, which possibly participates in the reproductive propagation of PrP^Sc.

MATERIALS AND METHODS

Expression of mouse PrP in a baculovirus-insect cell system.

A DNA fragment encoding amino acid residues 1 to 254 of mouse PrP was amplified by PCR from mouse brain cDNA using the forward primer 5′-CGGGATCCGCCACCATGGCGAACCTTGGCTAC-3′ and reverse primer 5′-CCCAAGCTTTCATCCCACGATCAGGAAG-3′. The PCR fragment was cloned into pFastBac1, a baculovirus transfer vector (Invitrogen, Carlsbad, CA), at BamHI and HindIII sites according to the manufacturer's protocol (Invitrogen). pFastBac1 with no insert was used as the control construct. The initial viral stocks were obtained by transfecting Spodoptera frugiperda 21 (Sf21) cells with a bacmid DNA-Cellfectin (Invitrogen) mixture, and their titers were improved through serial infections. HighFive cells (Invitrogen) were infected with a third viral stock and harvested for 72 h at 27°C. The harvested cells were used as the PrP^C substrate for the PMCA assay. The yield of Bac-PrP in the insect cells was approximately 16 μg/10⁶ cells. GPI-anchorless Bac-PrP was prepared by replacing the codon AGC of Ser 231 with the stop codon TGA. A DNA fragment encoding amino acid residues 1 to 230 of mouse PrP^C was amplified by PCR from mouse brain cDNA using the forward primer 5′-CGGGATCCGCCACCATGGCGAACCTTGGCTAC-3′ and reverse primer 5′-CCCAAGCTTTCAGGATCTTCTCCCGTCGTAATAG-3′. Expression of GPI-anchorless Bac-PrP was performed as well as that of wild-type Bac-PrP.

Phosphatidylinositol-specific phospholipase C digestion and TX-114 phase partitioning.

Triton X-114 (TX-114) phase partitioning analysis was performed essentially as described previously (3). To detach Bac-PrP from the cell surface by removing the lipid portion from the GPI anchor, phosphatidylinositol-specific phospholipase C (PIPLC) digestion was performed on intact Bac-PrP-expressing insect cells and mouse PrP^C-overexpressing N2a cells by immersion in phosphate-buffered saline (PBS) containing 0.2 U PIPLC/ml at 37°C for 1 h. To remove detached cells and debris, the medium of PIPLC-treated cells was centrifuged at 3,000 × g for 3 min and then centrifuged at 20,000 × g for 3 min. PrP^C in the media from PIPLC-treated cells and nontreated cells was analyzed by adding 0.25 volume of TX-114 stock (3% TX-114 in PBS) to the media at 4°C. PrP^C included in cells was analyzed by directly extracting the nontreated cells in PBS containing 0.25 volume of TX-114 stock at 4°C. Each sample was centrifuged at 3,000 × g for 3 min at 4°C to remove any insoluble aggregates. The supernatants were heated at 37°C for 3 min to induce phase transition and centrifuged at 3,000 × g for 3 min at 25°C. The upper (aqueous) phase was again mixed with 0.25 volume of TX-114 stock at 4°C and then heated. The second upper phase was added to the first one. Proteins in the upper and lower (detergent) phases were precipitated with 5 volumes of methanol, and PrP^C in each phase was detected by Western blot (WB) analysis. TX-114 phase partitioning of E. coli recPrP was also performed under the same conditions described above.

Preparation of brain homogenate.

To avoid contamination, normal ICR mouse and Prnp^0/0 mouse (37) BHs were prepared in a laboratory in which infected materials had never been handled. Brains of ICR and Prnp^0/0 mice were homogenized at a 20% (wt/vol) concentration in PBS containing a complete protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). The BHs were stored at −80°C until further use. The BHs were mixed with an equal volume of 2× PMCA buffer (PBS containing 2% Triton X-100 and 8 mM EDTA), and the 10% (wt/vol) BHs were subjected to PMCA. PrP^Sc seeds used in the present study were the mouse-adapted scrapie strains: Chandler, 79A, 22L, ME7, Obihiro (32), and Tsukuba-2 (unpublished data). These prion strains were propagated in ICR mice. The brains of mice at the terminal stage of the disease were homogenized at a 10% concentration (wt/vol) in PBS. All animal experiments were performed in accordance with National Institute of Animal Health guidelines.

PMCA.

PMCA was carried out using the automatic cross-ultrasonic protein-activating apparatus (Elestein 070-GOT; Elekon Science Corp., Chiba, Japan) as previously reported (26). Amplification was performed with 40 cycles of sonication (pulse oscillation for 3 s was repeated five times at 0.1-s intervals) followed by incubation at 37°C for 30 min with gentle agitation. For Bac-PrP-based amplification, the PrP^C substrate was prepared by adding 5 μl of Bac-PrP virus-infected cell suspension (5 × 10⁴ cells/μl) to 95 μl of 10% Prnp^0/0 BH. The amount of Bac-PrP in each reaction mixture (total, 100 μl) was approximately 4 μg (40 μg/ml), which was about 1.6 times higher than that of brain-PrP^C in 100 μl of normal ICR BH. For amplification using normal BH as the PrP^C substrate (BH-PMCA), 100 μl of 10% ICR mouse BH was used. The amplified product obtained after the first round of amplification was diluted 1:10 with the PrP^C substrate, and a second round of amplification was performed. This process was repeated when necessary.

Western blotting.

The PMCA products were digested with 50 μg/ml proteinase K (PK) at 37°C for 1 h. An equal volume of 2× SDS sample buffer was added to the samples and boiled for 5 min. The samples were separated by SDS-PAGE in NuPAGE 12% Bis-Tris gels (Invitrogen) and transferred onto polyvinylidene difluoride (PVDF) membranes. The membranes were typically probed with anti-PrP horseradish peroxidase-conjugated monoclonal antibody T2 (15), which recognizes the discontinuous epitope of amino acid residues 132 to 156 of mouse PrP. The SAF32 antibody (SPI-Bio, Montigny le Bretonneux, France) was used for the detection of full-length PrP^C. The blotted membrane was developed with SuperSignal West Dura extended-duration substrate (Pierce, Rockford, IL), and chemiluminescence signals were detected using Chemiimager (Alpha InnoTec, San Leandro, CA).

Expression of Escherichia coli recombinant PrP.

A DNA fragment encoding amino acid residues 23 to 230 of mouse PrP was amplified by PCR from mouse brain cDNA using the forward primer 5′-TCTAGCTGTCATATGAAAAAGCGGCCAAAGCCTG-3′ and reverse primer 5′-AGCTGTAAGCTTTCATCAGGATCTTCTCCCGTCGTAATAG-3′. The PCR fragment was cloned into pET11a (Invitrogen, Carlsbad, CA) at NdeI and HindIII sites. The vector was transformed into E. coli BL21(DE3)RIL (Stratagene, La Jolla, CA), and nontagged full-length mouse PrP was expressed as an inclusion body. The cells were lysed with Bugbuster reagent (Novagen, Madison, WI) according to the manufacturer's instructions, and the inclusion bodies were solubilized in 6 M guanidine hydrochloride (GdnHCl), 0.5% Triton X-100, 10 mM 2-mercaptethanol, and 100 mM sodium phosphate in 10 mM Tris-HCl (pH 8.0). The lysate was loaded onto a Co²⁺ chelating column (HisTALON Superflow Cartridge; Clontech, Mountain View, CA). E. coli recPrP was refolded according to the method of Atarashi et al. (2) and was eluted with 500 mM imidazole, 100 mM sodium phosphate, and protease inhibitor in 10 mM Tris-HCl (pH 5.8). E. coli recPrP was finally dialyzed in PBS.

PK degradation assay.

To compare resistance of PrP^res to PK digestion and GdnHCl denaturation, the concentrations of the amplified products obtained after 10 rounds of serial PMCA and the infected BH were adjusted with 10% Prnp^0/0 BH to obtain similar signal intensities of PrP^res in WB analysis. These samples were digested with various concentrations of PK (50 to 5000 μg/ml) at 37°C for 1 h. Boiling in 1× SDS sample buffer stopped the enzymatic digestion. The samples were analyzed by WB in three independent experiments, and the average intensity of the PrP^Sc signal in each sample was expressed as a percentage relative to that in the sample digested with 50 μg/ml of PK. The results were analyzed by one-way analysis of variance (ANOVA) and the Tukey-Kramer multiple-comparison test. The PK₅₀ value, which is the concentration of PK needed to reduce the signal intensity by half, was estimated from an approximate curve calculated from the experimental data.

GdnHCl denaturation assay.

The PMCA products and the infected BH, which were prepared in approximately the same concentration of PrP^res, were incubated with agitation at room temperature for 2 h in various concentrations of GdnHCl (0 to 4 M). The samples were further incubated with agitation at 4°C for 30 min in the presence of 10% Sarkosyl and then centrifuged at 100,000 × g for 1 h. The pellets were dissolved directly in PK solution (50 μg/ml) and incubated at 37°C for 1 h. WB in three independent experiments was performed to analyze the samples, and the average intensity of the PrP^Sc signal in each sample was expressed as a percentage relative to that in the sample prepared without GdnHCl. The results were analyzed statistically as described above.

Bioassay.

The PMCA products subjected to bioassay were prepared as follows. Chandler-infected BH was diluted 1:1,000, and then one round of PMCA was performed. Next, the process of 1/10 dilution of the PMCA product and its subsequent amplification was repeated three times. After that, the process of one-fifth dilution of the PMCA product and its subsequent amplification was repeated six times. The PMCA products were diluted 1:10 with PBS, and 20 μl of each diluted sample was injected intracerebrally into 3-week-old C57/BL6J female mice. Because ICR mice used for the propagation of the Chandler prion strain often develop hydrocephalus after inoculation of the PMCA products, C57/BL6J mice were used for the bioassay in the present study. Densitometric analysis of WBs revealed that PrP^res signal intensities in the Bac-PrP-based and BH-based PMCA products were approximately 40- and 30-fold lower, respectively, than those in an equal volume of 10% infected BHs used as the PrP^Sc seed. Therefore, Chandler-infected BHs were diluted 1:400 and 1:300 and injected into the mice, and the mice were used as the controls. Animals at the terminal stage of the disease were sacrificed, and the left hemisphere of the brain was stored at −80°C for biochemical analysis. The average survival times of the experimental groups were analyzed by one-way ANOVA and the Tukey-Kramer multiple-comparison test.

Histopathological studies.

The right hemisphere was fixed in 10% buffered formalin solution. Coronal slices of the brain were cut, immersed in 98% formic acid to reduce infectivity, and then embedded in paraffin wax. Sections with a thickness of 4 μm were cut and stained with hematoxylin and eosin (HE) or analyzed by immunohistochemistry. For the 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 (12). For immunohistochemistry, PrP^Sc was detected in brain sections by the hydrated autoclaving method using anti-PrP monoclonal antibody 31C6 against the epitope of amino acids 143 to 149 of the mouse prion protein (18). Immunoreactions were developed using anti-mouse universal immunoperoxidase polymer [Nichirei Histofine Simple Stain MAX-PO (M); Nichirei, Tokyo, Japan] as the secondary antibody and 3,3′-diaminobenzidine tetrachloride as the chromogen.

RESULTS

Expression of mouse recombinant PrP in a baculovirus-insect cell system.

To confirm the expression of recPrP in insect cells, the virus-infected cells were solubilized and analyzed by WB (Fig. 1 A). Three distinct bands were recognized in the lysate of the virus-infected cells, but no signal was detected in the lysate prepared from mock-infected cells. The molecular masses of the three bands of baculovirus-derived PrP (Bac-PrP) were 31.1, 29.5, and 27.2 kDa, respectively, and the banding pattern was different from that of brain-derived PrP^C (brain-PrP^C). To examine whether Bac-PrP was modified properly with N-glycans as observed in the brain-PrP^C, WB was performed on tunicamycin-treated cell lysates. A single band with a molecular mass close to that of the 27.2-kDa band of Bac-PrP from untreated cells was detected in the lysate, suggesting that Bac-PrP was N glycosylated. In non-tunicamycin-treated lysates, three types of Bac-PrP (di-, mono-, and unglycosylated isoforms) were detected in the insect cells (Fig. 1B).

FIG. 1.

Expression of mouse PrP in insect cells. (A) WB analysis of lysates of mock-infected insect cells (negative control [NC]), Bac-PrP-expressing insect cells (Bac), and normal mouse brain homogenates (BH). SAF32 antibody was used for the detection of PrP. (B) Banding patterns of Bac-PrP from tunicamycin-treated (+) and nontreated (−) cells. The cells were exposed to 10 μg/ml of tunicamycin for 2 days after incubation with the constructed virus for 1 day. (C) Triton X-114 (TX-114) partitioning of Bac-PrP. Bac-PrP-expressing insect cells and PrP^C-overexpressing N2a cells were incubated in PBS in the presence (+) or absence (−) of 0.2 U PIPLC/ml at 37°C for 1 h. The medium from each cell culture was partitioned in TX-114 (lanes 1 to 4 and 7 to 10). Nontreated insect cells (lanes 5 and 6), N2a cells (lanes 11 and 12), and E. coli recPrP (lanes 13 and 14) were extracted in TX-114 at 4°C, and the supernatants obtained by centrifugation were partitioned with TX-114. WB was performed to analyze the resultant aqueous (lanes A) and detergent-rich (lanes D) phases.

To investigate whether Bac-PrP was anchored by GPI to the cell surface, we performed Triton X-114 phase partitioning on the lysates of insect cells expressing Bac-PrP and on the medium from PIPLC-treated insect cells (Fig. 1C). Almost all Bac-PrP from nontreated cell lysates was partitioned into the detergent phases (D), as was PrP^C from nontreated N2a cell lysates (lanes 5, 6, 11, and 12). This is in contrast to the case for E. coli recPrP, which was almost entirely partitioned into the aqueous phase (A) (lanes 13 and 14). When Triton X-114 partitioning was applied to PIPLC-treated cell culture medium, Bac-PrP was liberated from cell membranes in the presence of PIPLC and partitioned mainly into the aqueous phase, as also observed for PrP^C in PIPLC-treated N2a cell culture media (lanes 1, 2, 7, and 8). In contrast to this, liberated Bac-PrP and brain-PrP^C were rarely detected in both aqueous and detergent phases from the medium of nontreated cells (lanes 3, 4, 9 and 10). These results strongly suggest that Bac-PrP is GPI anchored to the plasma membrane, as is brain-PrP^C.

Generation of PK-resistant Bac-PrP.

Bac-PrP-expressing insect cells were mixed with either PMCA buffer or 10% Prnp^0/0 BH, and these cell suspensions were used as the PrP^C substrate. PrP^Sc from Chandler- and ME7-infected BHs were used as PrP^Sc seeds. Generation of proteinase K-resistant PrP (PrP^res) was hardly detected when amplification was performed in PMCA buffer (Fig. 2 A, lanes 1 to 4), as observed with the negative controls (lanes 9 to 12). Since RNA and/or lipids are reported to be effective at stimulating the generation of PrP^Sc by PMCA (11, 35), we examined the effect of these substances on Bac-PMCA. However, no significant amplification of PrP^res was recognized in any of the samples examined using the present experimental conditions (Fig. 2B). In contrast to these observations, Bac-PrP was converted into PrP^res when amplification was performed in the presence of 10% Prnp^0/0 BH. This finding is significant for the generation of PrP^res, because non-N-glycosylated, GPI-anchorless bacterial recPrP does not show any notable ability to convert into PrP^res in PMCA buffer even in the presence of Prnp^0/0 BH (Fig. 3 A). We further examined the significance of GPI anchoring and N-linked glycosylation of Bac-PrP in the amplification by PMCA using Bac-PrP mutants. As reported for the amplification of mammalian PrP^Sc (20), GPI anchoring of Bac-PrP was necessary to generate PrP^res, because the GPI-anchorless Bac-PrP mutant was not useful as a PrP^C substrate even in the presence of Prnp^0/0 BH (Fig. 3B). In addition, wild-type Bac-PrP prepared from PIPLC-treated insect cells was useless as a substrate (see Fig. S1 in the supplemental material). On the other hand, the non-N-glycosylated Bac-PrP mutant retained its amplification ability (see Fig. S2 in the supplemental material). These results suggest that Prnp^0/0 BH and Bac-PrP modified with a GPI anchor were necessary for the generation of Bac-PrP^res by PMCA.

FIG. 2.

In vitro amplification of mouse-adapted scrapie PrP^Sc by Bac-PMCA. (A) PMCA was performed using Bac-PrP-expressing cells (Bac-PMCA) or normal mouse BH (BH-PMCA) as a PrP^C substrate. Bac-PMCA was performed in the absence (PMCA buffer only [Buf]) or presence of 10% Prnp^0/0 BH. Homogenates (10%) of Chandler (Chan)- and ME7-infected brains were diluted 1:1,000 with each PrP^C substrate. Amplification was carried out in duplicate except for the negative controls. The PMCA products were digested with 50 μg/ml of PK at 37°C for 1 h and analyzed by WB. (B) Amplification of Chandler PrP^Sc (diluted 1:1,000) by Bac-PMCA was performed in the presence of Prnp^0/0 BH, poly(A) in PMCA buffer, and PMCA buffer alone (−). Amplification was also performed in the presence of poly(A) and synthetic lipid (POPG) as described by Wang et al. (35) and in the TX100-SDS buffer used for bacterial recPrP amplification (2).

FIG. 3.

Effect of the GPI anchor in Bac-PMCA. (A) Amplification of Bac-PrP and E. coli-derived recPrP in the presence of Prnp^0/0 BH. Purified E. coli recPrP (4 to 10 μg) was used as the PrP^C substrate, and PMCA was performed under the same conditions as for Bac-PMCA. WB analysis was performed after PK digestion. (B) Amplification of GPI-anchorless Bac-PrP. Insect cells (2.5 × 10⁵ cells) containing about 4 μg of the mutant PrP were used as the PrP^C substrate, and serial PMCA (R1 to R4) was performed in the presence of Prnp^0/0 BH. WB analysis was performed before and after PK digestion. GPI-anchorless PrP molecules were retained in the cells (data not shown), probably due to a defect in intracellular transport, and were therefore not N glycosylated (lower panel).

Availability of Bac-PrP for amplification of various prion strains.

We compared the amplification efficiencies of PrP^Sc of various prion strains in Bac-PrP- and BH-based amplifications (Fig. 4 A). Among six mouse-adapted scrapie prion strains, Chandler-, 79A-, and 22L-derived PrP^Sc could be amplified during three rounds of amplification by serial Bac- and BH-PMCA. However, the amplification efficiencies for PrP^Sc from the ME7 and Obihiro strains were insufficient to maintain both Bac- and BH-PrP^res signals in WB analysis, and Tsukuba-2 PrP^Sc was hardly amplified by both PMCAs. These observations suggest that Bac-PrP and brain-PrP^C have similar capabilities for the amplification of PrP^Sc from these strains in vitro.

FIG. 4.

Amplification capability and efficiency of Bac- and BH-PMCA. (A) Six mouse-adapted scrapie strains were used as the PrP^Sc seed. Homogenates (10%) of each prion strain-infected brain were diluted 1:1,000 with Prnp^0/0 BH containing Bac-PrP-expressing cells or 10% normal BH, and three rounds of serial PMCA were performed. WB analysis was performed after PK digestion. Lanes: a, no-seed control; b, Chandler; c, 79A; d, 22L; e, ME7; f, Obihiro; g, Tsukuba-2. (B) Sensitivities of detection of Chandler PrP^Sc by Bac- and BH-PMCA. Chandler-infected 10% BH was diluted 10⁻³ to 10⁻¹¹ with each PrP^C substrate, and then six rounds of serial PMCA were performed. WB analysis was performed following PK digestion. −, no-seed control.

We further investigated the sensitivity of detection of PrP^Sc by Bac-PMCA. Chandler PrP^Sc in the 10% infected BH diluted 10⁻¹⁰ was detected after six rounds of Bac-PMCA, whereas PrP^Sc was detected in the samples diluted 10⁻⁸ after four rounds of BH-PMCA (Fig. 4B). Therefore, Bac-PMCA was more sensitive than BH-PMCA in the detection of Chandler PrP^Sc.

Biochemical properties and infectivity of Bac-PrP^res.

We compared the biochemical properties of Bac-PrP^res with those of the BH-PrP^res and Chandler PrP^Sc seeds. The PK₅₀ value of Bac-PrP^res was 220 μg/ml, whereas BH-PrP^res and the PrP^Sc seed were more resistant to PK digestion, with PK₅₀ values of 1,026 and 1,228 μg/ml, respectively (Fig. 5 A). In contrast, the GdnHCl₅₀ values of Bac-PrP^res were similar to those of BH-PrP^res and the PrP^Sc seed (Fig. 5B). Infectivity of Bac-PrP^res was examined by inoculating C57BL6J mice with the PMCA products (Table 1). The mice inoculated with Bac-PrP^res and BH-PrP^res developed typical clinical signs of prion disease, such as a rough coat, hunchback, weight loss, and ataxia, and they died after average periods of 193 ± 3 and 190 ± 3 days, respectively. Control mice inoculated with Chandler-infected BH diluted 1:400 and 1:300 died after average periods of 182 ± 3 days and 178 ± 4 days, respectively. There was no significant difference between the survival periods of the PMCA product-inoculated mice and the control mice.

FIG. 5.

Biochemical properties of PMCA products generated by Bac- and BH-PMCA and the Chandler PrP^Sc seed. Bac- and BH-PrP^res obtained after 10 rounds of serial PMCA and Chandler PrP^Sc were prepared to obtain similar concentrations for the PK degradation assay (A) and the GdnHCl denaturation assay (B). WB was performed to analyze the samples from five independent experiments. The average relative intensities of the PrP^res signal with standard errors (SE) at each concentration of PK and GdnHCl are represented graphically. Asterisks denote significant differences among the samples (P < 0.05).

TABLE 1.

Mean survival times of C57/BL6J mice following intracerebral inoculation

Inoculuma	Transmission rate, % (no. dead/total no.)	Mean survival time ± SE (days)
Bac-PMCA
PMCA +, seed +	100 (16/16)	193 ± 3
PMCA −, seed +	0 (0/6)	>370
PMCA +, seed −	0 (0/3)	>270
Chandler BH (1/40 dilution)	100 (6/6)	182 ± 3
BH-PMCA
PMCA +, seed +	100 (12/12)	190 ± 3
PMCA −, seed +	0 (0/6)	>370
PMCA +, seed −	0 (0/3)	>270
Chandler BH (1/30 dilution)	100 (8/8)	178 ± 4

^aThe final dilution of the Chandler seed in the “seed +” samples was 6.4 × 10⁻¹¹. The 10% Chandler-infected BH was diluted 1:40 or 1:30 to obtain a PrP^res level similar to that for the PMCA products in WB analysis. The samples were diluted 1:10 in PBS and intracerebrally inoculated into the mice.

Histopathological analysis of Bac-PrP^res-inoculated mice.

To characterize the neuropathological properties of the Bac-PMCA products, we examined the regional profiles of neuronal vacuolation scores in the brains of the affected mice (Fig. 6 A). No significant difference in the vacuolation scores in brain regions of the mice inoculated with the Bac- and BH-PMCA products and the Chandler-infected BH was observed, except for the septal nuclei of the paraterminal body. The immunohistochemical results for the occipital cortex from infected mice are shown in Fig. 6B. In the mice inoculated with the PMCA products and Chandler-infected BH, PrP^Sc accumulated over the whole region of the brain sections with a diffuse distribution and synaptic-like immunostaining, and the accumulation patterns of PrP^Sc were very similar. These results suggest that the particular neuropathological properties of the Chandler PrP^Sc seed were mostly preserved in Bac-PrP^res, as observed for BH-PrP^res.

FIG. 6.

Histopathological analysis of the brains of mice inoculated with Bac- and BH-PMCA products and Chandler-infected BH. (A) Vacuolation profile in different brain areas of C57BL/6J mice inoculated with Bac-PMCA products (C57/Bac-PMCA), BH-PMCA products (C57/BH-PMCA), and Chandler BH (C57/Chandler BH). Brain regions are as follows: 1, dorsal medulla; 2, cerebellar cortex; 3, superior collicullus; 4, hypothalamus; 5, thalamus; 6, hippocampus; 7, septal nuclei of the paraterminal body; 8, cerebral cortex at the levels of 4 and 5; 9, cerebral cortex at the level of 7. The average lesion scores (n = 5 to 10 animals per group) and SE are shown in the graph. Asterisks denote significant differences among the samples (P < 0.01). (B) Vacuolation (upper panels) and PrP^Sc accumulation (lower panels) in the brains of C57/Bac-PMCA, C57/BH-PMCA, and C57/Chandler BH mice. Mice inoculated with the products obtained after 13 rounds of unseeded Bac-PMCA were used as the control. Control mice did not show any clinical signs of the disease and were killed at 245 days after inoculation.

Biochemical properties of Bac-PrP^Sc propagated in vivo.

We further examined the WB profiles, PK resistance, and GdnHCl denaturation of PrP^Sc accumulated in the brains of C57BL mice inoculated with the Bac- and BH-PMCA products. Chandler-infected BH was also inoculated intracerebrally into C57BL mice as the control PrP^Sc. The glycosylation patterns and molecular masses of Bac-PrP^Sc were indistinguishable from those of BH-PrP^Sc, Chandler-derived PrP^Sc, or the original Chandler PrP^Sc (Fig. 7 A). In addition, no significant differences were observed in the PK sensitivities or GdnHCl denaturation abilities of Bac-PrP^Sc, BH-PrP^Sc, or Chandler-derived PrP^Sc (Fig. 7B).

FIG. 7.

Biochemical properties of PrP^Sc accumulated in the brains of C57BL/6J mice inoculated with Bac- and BH-PMCA products and Chandler BH. (A) WB analysis of PrP^Sc in the brains of C57/Bac-PMCA, C57/BH-PMCA, and C57/Chandler BH mice (three each). WB analysis was performed after PK digestion. WB profiles of PrP^Sc that accumulated in these mice were very similar and were indistinguishable from that of the original Chandler PrP^Sc propagated in ICR mice. (B) Biochemical properties of each PrP^Sc were compared using the PK degradation assay (left panel) and the GdnHCl denaturation assay (right panel). WB was performed to analyze the samples from four independent experiments. The average relative intensities of the PrP^res signal with SE at each concentration of PK and GdnHCl are represented graphically. No significant differences were observed among the PrP^Sc samples.

DISCUSSION

In this study, we succeeded for the first time in generating PrP^Sc by PMCA using recombinant PrP produced in the baculovirus-insect expression system, and we found that Bac-PrP could be amplified by PMCA in the presence of the Prnp^0/0 BH. Although the Bac- and BH-PMCA products showed different WB profiles (Fig. 2A) and PK sensitivities (Fig. 5A), probably due to the inherent posttranslational modifications of PrP processed in insect and mammalian cells, the Bac- and BH-PMCAs were similar in their amplification capabilities with several scrapie prion strains and efficiencies of amplification of Chandler PrP^Sc. Furthermore, using Bac-PrP mutants, we revealed that Bac-PrP modification with the GPI anchor was required for amplification in vitro. In particular, PrP^Sc that had accumulated in the brains of mice inoculated with the Bac-PMCA product strikingly resembled in its biological properties PrP^Sc in brains inoculated with the BH-PMCA product or Chandler.

Since the GPI anchor structure in insects has not been elucidated (27), details of the GPI anchor of Bac-PrP remain unknown. However, Bac-PrP appeared to function as a PrP^C substrate as well as brain-PrP^C. In the present study, insect cells expressing Bac-PrP were directly used for the PrP^C substrate for convenience. We partially purified Bac-PrP from the lysate of insect cells and confirmed that such Bac-PrP could be amplified by PMCA in the presence of Prnp^0/0 BH at a level equivalent to the amplification achieved using intact cells (see Fig. S3 in the supplemental material). Therefore, Bac-PMCA is not very susceptible to admixtures contained in the insect cell lysate.

In vitro studies using the E. coli-derived recPrP and PMCA technique have demonstrated that recPrP^Sc could be amplified using the PrP^Sc core prepared from infected BH as the seed in the reaction mixture, which contained only salt and detergent (19), and that de novo generation of recPrP^Sc was stimulated in the presence of RNA and lipids (35). These finding suggest that posttranslational modification of PrP is not necessary for the amplification and generation of recPrP^Sc by PMCA. In contrast to these results, the present study indicates that Bac-PrP modification with a GPI anchor was essential for the efficient amplification of Bac-PrP^res. Most GPI-anchorless Bac-PrP molecules were retained in the cells (data not shown), probably due to a defect in intracellular transport, and therefore were not N glycosylated (Fig. 3B, lower panel). However, N glycosylation of PrP^C itself is not involved in the generation of PrP^Sc, because non-N-glycosylated but GPI-anchored Bac-PrP^res could be amplified by PMCA and such PMCA products were infectious to wild-type mice (see Fig. S2 in the supplemental material). Therefore, two different stimulation pathways responsible for the generation of PrP^Sc can be considered: one is a GPI anchor-independent pathway and the other is a GPI anchor-dependent pathway. The importance of the latter pathway in the pathogenesis of prion disease has been suggested in an in vivo study using GPI-anchorless PrP-expressing mice (6).

A GPI anchor is inserted into the lipid raft domains of the cell membrane, and therefore PrP^C can interact with cell membrane components. A number of studies have demonstrated that various cell-derived molecules possibly participate in the generation of PrP^res or PrP^Sc. For example, these include polyanionic molecules such as nucleic acids (8, 11, 35) and sulfated glycans (36), lipids (17), molecular chaperone proteins such as heat shock protein 60 (hsp60) and hsp104 (10), and PrP-interacting molecules such as laminin receptor (22). Recently, some factors that are contained in lipid raft fractions from mammalian species have been shown to be necessary for BH-PMCA (1). Therefore, some kind of cofactors could facilitate the propagation of PrP^Sc through interaction with the GPI anchor itself or GPI-anchored PrP^C. Since E. coli-derived recPrP^Sc (19) and brain-derived PrP^Sc generated in the absence of BH (11) showed biological properties different from those of the PrP^Sc seed, interaction between GPI-anchored PrP^C and cofactors contained in BH appears to be necessary for reproductive propagation of PrP^Sc, although such stimulation pathways might not necessarily be needed in order to trigger the generation of PrP^Sc.

In the present study, we demonstrated that recombinant PrP modified with an insect cell-type GPI anchor was functional as a PrP^C substrate for the amplification of mammalian PrP^Sc, and this amplified product showed properties very similar to those of the PrP^Sc seed in its pathogenicity in wild-type mice and in the formation of neuropathological lesions in infected mice. The great advantage of the Bac-PMCA technique is that Bac-PrP can be prepared free from contamination of any cofactors contained in mammalian cells. Therefore, this Bac-PMCA technique that we developed here will be a powerful tool for elucidating GPI anchor-mediated stimulation pathways and for identifying the cofactors required for reproductive propagation of PrP^Sc.