Introduction

Prions are infectious pathogens that cause a group of transmissible prion diseases in animals and humans. Currently there is no cure for prion diseases, largely because the molecular mechanism underlying prion formation is poorly understood. The scrapie isoform (PrPSc) of the cellular prion protein (PrPC) is the only known component of prions. The conversion of PrPC into PrPSc constitutes a key molecular event in the pathogenesis of prion diseases; however, the mechanism underlying the conversion remains unclear. It has been proposed that prion formation occurs in a template-assisted process involving the physical interaction of the PrPSc template and the PrPC substrate1. Indeed, early in vivo studies indicated that interaction between non-homologous PrP molecules inhibits the disease process2,3,4. The incorporation of chimeric PrP into PrPSc was influenced by the PrP sequence in scrapie-infected cell lines expressing chimeric mouse-hamster PrP5. Subsequently, Priola et al provided direct evidence that heterologous PrP molecules, which differed by as little as one residue, interfere with the generation of PrPSc in scrapie-infected mouse cells (ScN2a)6. Based on this result, as well as previous studies, the authors proposed three possible mechanims for the interference. First, interaction between dissimilar PrPSc and PrPC molecules might slow the aggregation and accumulation of PrPSc by interfering with the interaction of similar PrP monomers7,8,2. Second, incorporation of non-homologous PrP molecules into PrPSc aggregates might lead to a destabilization of the aggregates9. Finally, exogenous PrP molecules might inhibit the interaction of the endogenous PrP with cellular ligands10.

Studies with transgenic mice expressing human or mouse/human chimeric PrP implied that a species-specific cofactor, termed protein X, is necessary for PrPSc formation11. Four mouse specific substitutions in the C-terminal region of PrP, including residues 167, 171, 214 and 218 were identified that inhibit the conversion of wild-type PrPC in a dominant-negative manner in scrapie-infected cells12. These residues were proposed to form a discontinuous epitope that interacts with protein X. However, although many putative protein X genes have been proposed, knockout of these genes in mice failed to significantly alter incubation times13. Moreover, the recombinant Q218K variant, one of the four dominant negative mutants, inhibited the polymerization of recombinant wild-type PrP in the absence of protein X14. The dominant-negative effect observed in pure recombinant molecules was presumably mediated by physical interaction between the Q218K variant and wild-type PrP. Using the protein misfolding cyclic amplification (PMCA) assay with wild-type and mutant PrP expressed in Chinese hamster ovary cells as substrates, Geoghegan et al further demonstrated that trans-dominant inhibition of prion propagation in vitro was not mediated by an accessory cofactor and proposed that PrP molecules compete for binding to a nascent seeding site on newly formed PrPSc molecules15.

In the current study, we demonstrate that unglycosylated and anchorless recombinant full-length human PrP23-231 is able to dramatically inhibit human PrPSc amplification in vitro. Moreover, this inhibition also occurs in a scrapie-infected cell model. Although to a lesser extent, recombinant PrP from other species also inhibits human PrPSc amplification. We show that the inhibition may depend on direct interaction of the inhibitory recombinant PrP with human PrPSc using a capture strategy.

Results

Amplification of human PrPSc is inhibited by unglycosylated and anchorless recombinant human PrP

A recent study in our lab suggested that the glycoform-selective prion formation observed in unique sporadic and familial forms of prion disease may involve changes in N-linked glycosylation16. Indeed, using the serial PMCA, Nishina et al. observed that the formation of Sc237 hamster prions was dependent on substoichiometric levels of unglycosylated PrPC molecules isolated from the hamster brain17. Moreover, recombinant hamster PrP that lacks both glycans as well as the glycophosphatidylinositol (GPI) anchor was found to inhibit amplification of hamster PrPSc18,19 in a standard PMCA reaction. To investigate the effect of unglycosylated and anchorless PrP on human PrPSc formation, we performed the PMCA assay in which human PrPSc from brain homogenates of an iatrogenic CJD (iCJD) was used as the seed while human PrPC from brain homogenates of transgenic mice expressing wild-type human PrP-129V was used as the substrate. The unglycosylated and anchorless recombinant full-length human PrP23-231 (rHuPrP23-231) with methionine at the polymorphic residue 129 was added into the PMCA. In controls lacking rHuPrP23-231, intense PK-resistant PrPSc (PrPres) bands were detected in the sample subjected to PMCA while virtually no PrP was detectable in the non-PMCA sample, suggesting that significant amplification of PrPSc occurred (Figure 1A, lanes 1 and 2). In contrast, in the sample that contained no PrPSc seeds, PK-resistant PrP (PrPres) was not detectable (Figure 1A, lanes 7 and 8). In the presence of 0.2 μM rHuPrP23-231, virtually no PrPres bands were detectable in the sample subjected to PMCA, similar to the non-PMCA sample, indicating that rHuPrP23-231 inhibited the amplification of PrPSc. In order to assess whether other proteins known to interact with PrP could inhibit the amplification of PrPSc, we used human protein disulfide isomerase (PDI)20. When PMCA was performed in the presence of the same amount of recombinant PDI (rPDI), intense PrPres bands were still detected, similar to the sample without any recombinant proteins. Thus, the inhibition is rHuPrP-specific. To investigate whether the methionine (M) or valine (V) polymorphism at residue 129 affects the inhibition efficiency, we replaced the 129M-rHuPrP with an equal amount of 129V-rHuPrP in the PMCA reaction. No significant difference in the inhibition efficiency was observed (data not shown). As a result, in subsequent experiments we continued using 129M-rHuPrP. Our previous study found that an anti-tumor drug mechlorethamine (MCT) inhibited hamster PrPSc 263K amplification using PMCA21. Here we determined the effect of MCT on human PrPSc amplification by using the same substrates and seeds in the absence or presence of MCT. The intensity of PK-resistant PrP was significantly decreased in the sample containing MCT compared to the sample without MCT (Figure 1B). Thus, we confirmed that MCT also inhibits amplification of human PrPSc (Figure 1B).

Figure 1
figure 1

Inhibition of amplification of human PrPSc by recombinant human PrP (rHuPrP23-231) and mechlorethamine (MCT).

(A) Amplification of human PrPSc from iatrogenic CJD carrying valine (V)/valine polymorphism at residue 129 (129VV) of PrP characteristic of PrPSc type 2 (iCJDVV2, seeds) was carried out by PMCA in the presence of uninfected brain homogenates from humanized transgenic mice expressing PrP-129V (substrates). Lanes 1 and 2: Positive PMCA control without inhibitors; Lanes 3 and 4: PMCA in the presence of 0.2 μM of rHuPrP23-231 with methionine at polymorphic residue 129 (129M); Lanes 5 and 6: PMCA in the presence of 0.2 μM of recombinant human protein disulfide isomerase (rHuPDI); Lanes 1 through 6: PMCA with PrPSc seeds; Lanes 7 and 8: PMCA without PrPSc seeds and inhibitors; Lane 9: The PrP sample without PK-treatment; and Lanes 1 through 8 with PK-treatment. Samples without (−) or with (+) PMCA were treated with 100 μg/ml PK prior to SDS-PAGE and Western blotting with 3F4. Intense PrPSc is detectable in lane 2 (positive control) but not in lane 8 (negative control). However, amplification is virtually undetectable in the presence of rHuPrP23-231 (lane 4) while detectable in the presence of rHuPDI (lane 6). The blot is a representative of five independent experiments. (B) Amplification of human PrPSc from iCJD was carried out in the absence (lanes 3 and 4) or presence (lanes 5 and 6) of MCT. Amplification of PrPSc is inhibited in the presence of MCT (1.5 mM) compared to the sample in the absence of MCT. Lanes 3 through 6 were treated with PK while lanes 1 and 2 were not. The blot is a representative of three independent experiments.

Full size image

The effect of rHuPrP23-231 on PrPSc amplification is dose-dependent

To determine whether there is a dose-dependent effect on PrPSc amplification, we next conducted PMCA in the presence of different concentrations of rHuPrP23-231. Different amounts of rHuPrP23-231 ranging from 0 to 480 nM were added to the reaction containing both PrPSc seeds and PrPC substrates immediately before PMCA. The intensity of PrPres detected in samples subjected to PMCA was decreased as a function of the concentration of rHuPrP23-231 added (Figure 2A). The half maximal effective concentration (EC50) of the inhibition of PrPSc amplification by rHuPrP23-231 was approximately 60 nM (Figure 2B).

Figure 2
figure 2

Dose-dependent inhibition of PrPSc amplification by rHuPrP23-231.

(A) PMCA was performed with the mixture of human PrPSc (seeds) from iCJDVV2 and brain homogenates from TgWV (substrates) in the presence of different amounts of rHuPrP23-231 ranging from 0 to 480 nM. Samples without (−) or with (+) PMCA were subjected to PK-treatment prior to Western blotting with 3F4. (B) Percentage of inhibition of PrPSc amplification is a function of concentrations of rHuPrP23-231 added. The inhibition of PrPSc amplification by recombinant rHuPrP23-231 is dose-dependent and the half maximal effective concentration (EC50) is approximately 60 nM. The results are a representative of three independent experiments.

Full size image

Inhibition of PrPSc amplification by rHuPrP23-231 is species-specific

We then determined whether recombinant PrP from other species also inhibits human PrPSc amplification. The same concentration of recombinant human, mouse, bank vole and bovine PrP species was used. Compared to the PMCA sample in which no recombinant PrP was added, the samples that contained different recombinant PrP species exhibited varying degrees of inhibition. The efficiency of inhibition by the heterologous recombinant PrP species was much lower than that of the homologous rHuPrP23-231 (~20–30% vs ~100%) (Figure 3A and 3B). Therefore, the inhibition of PrPSc amplification by rHuPrP23-231 was species-specific. Indeed, although 200 nM recombinant mouse PrP23-231 (rMoPrP23-231) resulted in approximately 15% inhibition of human PrPSc amplification (Figure 3A and 3B), the same amount of rMoPrP23-231 inhibited more than 50% of mouse PrPSc amplification (Figure 3C and 3D). The EC50 of rMoPrP23-231 in the inhibition of mouse PrPSc (mouse 139A prion strain) was approximately 120 nM (Figure 3D). However, rHuPrP23-231 did not significantly inhibit amplification of mouse PrPSc in a standard PMCA reaction (Figure S1).

Figure 3
figure 3

Inhibition of PrPSc amplification by different species of recombinant PrP.

(A) PMCA was performed with the mixture of human PrPSc (seeds) from iCJDVV2 and brain homogenates from TgWV (substrates) in the presence of different species of recombinant PrP (0.2 μM each) including mouse (Mou: rMoPrP23-231 with 129M, home-made), human (Hum: rHuPrP23-231 with 129M), bank vole (BV: rBvPrP23-231 with 109I) and bovine (Bov: rBoPrP23-231 with 129M). (B) Inhibition of PrPSc amplification was quantified using densitometric analysis based on three independent experiments. Bars represent the percentage of all five amplified PrPSc with or without inhibitors to the amplified PrPSc without any inhibitors. Amplified PrPSc was calculated by subtracting the untreated PrP intensity (−) from the PMCA-treated PrP intensity (+) shown in (A). Of all recombinant PrP species examined, recombinant human PrP23-231 exhibited the highest inhibition compared to other species (**: p < 0.01; ***: p < 0.001). (C) PMCA was performed with mouse brain homogenates infected with prion 139A (seeds) and brain homogenates from wild-type mouse FVB (substrates) in the presence of different concentrations of the commercially-derived rMoPrP23-231 with 129M. (D) The inhibition of mouse prion 139A is dose-dependent and the half maximal effective concentration (EC50) is approximately 120 nM, which is based on three independent experiments.

Full size image

Effect of truncated PrP and PrPC- or PrPSc-specific binding reagents on human PrPSc amplification

We further determined which part of recombinant PrP is involved in the inhibition and investigated the effect of PrPC- or PrPSc-binding reagents on human PrPSc amplification. This included N-terminally-truncated recombinant human PrP90-231(Hu90), C-terminally-truncated recombinant human PrP23-145 (Hu145) and anti-PrP antibodies such as SAF32, 3F4, 6H4 and 8H4. We also investigated the effect of an anti-DNA antibody OCD4 and the gene 5 protein (g5p, a single stranded DNA-binding protein) that were previously shown to specifically bind to PrPSc but not to PrPC22. Compared to the PrPres intensity from the sample in the absence of recombinant PrP or antibodies, the PrPres intensity was decreased approximately 50% or more when rHuPrP90-231, rHuPrP23-145, g5p, or MCT were added to the reaction (p < 0.01 for Hu90, Hu145, or MCT; p < 0.001 for Hu23) (Figure 4A through 4D). PrPres was decreased approximately 10%–30% when SAF32, 3F4, 6H4, or OCD4 was added to the reaction, which was not statistically significant compared to the control containing no antibody (p > 0.05). A slight increase in the level of PrPres (~5%–10%) was observed in the sample containing the 8H4 antibody (p > 0.05). These results suggest that the inhibition of PrPSc amplification involves both N- and C-terminal domains of the protein. The SAF32, 3F4 and 6H4 antibodies that exhibited some inhibition are against residues 78–91, 106–112 and 145–152, respectively, while the 8H4 antibody that exhibited no inhibition is against residues 175–185 in human PrP. Again, rHuPrP23-231 (Hu23) virtually completely inhibited PrPSc amplification. It is most likely that the inhibitory sites are N-terminal to residue 175. Moreover, although both g5p and OCD4 specifically captured PrPSc, the efficiency of PrPSc inhibition was greater with g5p than with OCD4, suggesting that the two may have different binding sites on the PrPSc molecule. A second source of rHuPrP23-231 and rHuPrP90-231 generated previously23 was also examined and there were no significant differences in the inhibitory efficiency between currently and previously generated recombinant proteins (data not shown).

Figure 4
figure 4

Inhibition of human PrPSc amplification by truncated recombinant human PrP and different anti-PrP antibodies.

(A) PMCA was performed with the mixture of human PrPSc (seeds) from iCJDVV2 and brain homogenates from TgWV (substrates) in the presence of rHuPrP23-231 (Hu23), rHuPrP90-231 (Hu90), rHuPrP23-145 (Hu145), or different antibodies against PrP (0.1 μM each), respectively. The four antibodies include SAF32 against human PrP59-89, 3F4 against PrP105-112, 6H4 against PrP145-152 and 8H4 against PrP175-185. The result is a representative of three independent experiments. (B) Inhibition of PrPSc amplification was quantified using densitometric analysis based on three independent experiments. Of all recombinant PrP species and anti-PrP antibodies examined, recombinant human PrP23-231 exhibited the highest inhibition compared to others (**: p < 0.01; ***: p < 0.001). (C) PMCA was performed with the mixture of human PrPSc (seeds) from iCJDVV2 and brain homogenates from TgWV (substrates) in the presence of Hu23, Hu90, g5p, MCT, OCD4, or 3F4 (0.1 μM each), respectively. The Western blot shown is a representative of three independent experiments. (D) Inhibition of PrPSc amplification was quantified using densitometric analysis based on three independent experiments. In addition to rHuPrP23-231 and rHuPrP90-231, g5p and MCT also significantly inhibited PrPSc amplification (**: p < 0.01; ***: p < 0.001), whereas OCD4 and 3F4 did not (p > 0.05).

Full size image

Recombinant rHuPrP23-231 also inhibits murine PrPSc propagation in scrapie-infected mouse neuroblastoma cells (ScN2a)

We next determined whether rHuPrP23-231 is able to inhibit PrPSc propagation in scrapie-infected cells. Mouse prion-infected ScN2a cells were chosen in our study, due to the lack of human cell models of prion infection. This cell line is the most widely-used cell model, not only for studying the cell biology of prion replication, but also for screening therapeutic compounds24,25. ScN2a cells were incubated with concentrations of rHuPrP23-231 ranging from 0 to 1 μM for four days. While the cell toxicity, conducted as previously described26, was not observed in the cells incubated with rHuPrP23-231 at these concentrations, PrPres was decreased as a function of the increased concentration of recombinant PrP added to the media (Figure 5). No significant changes in the levels of β-actin were observed; β-actin was used to normalize the protein loading of the cell lysates.

Figure 5
figure 5

Inhibition of PrPSc propagation by recombinant human PrP23-231 in ScN2a cells.

Different amounts of rHuPrP23-231 ranging from 0 to 1 μM were added into the cell culture medium for four days. The cells were lysed and subjected to PK-digestion at 25 μg/ml prior to SDS-PAGE and Western blotting with 6D11. The intensity of PK-resistant PrPSc was significantly decreased at 0.1 μM of rHuPrP23-231 or greater. β-actin was determined to normalize the levels of individual samples examined. The blot is a representative of three independent experiments.

Full size image

Recombinant HuPrP23-231 binds to brain PrPSc but not PrPC

To determine whether the inhibition of PrPSc replication by rHuPrP23-231 results from its interaction with PrPC, PrPSc and/or both, we used a magnetic bead-based capture assay as previously described27. g5p- or PDI-beads were used as controls. No PrP was detected in the sample from the normal brain control after PK-treatment, although several bands were detected using the 3F4 antibody (Figure 6A). In contrast, a smear above 27 kDa was detected in the sample from sCJD; moreover, the three typical PrPres bands were detectable after PK-treatment (Figure 6A). Beads without any conjugated protein or antibody (empty) or conjugated with PDI did not capture PrP from either uninfected or prion-infected brain homogenates. As expected22,27, g5p-conjugated beads captured PrPSc (Figure 6A). To confirm that rHuPrP23-231 cannot capture PrPC, we incubated rHuPrP23-231-conjugated beads with normal brain homogenate or binding buffer alone. No differences were found between the two conditions (Figure 6B). The four major bands were monomeric or dimeric recombinant full-length or truncated PrP species, respectively. Therefore, our results indicate that recombinant PrP23-231, just like g5p, binds specifically to PrPSc, but not to PrPC.

Figure 6
figure 6

Binding of recombinant human PrP23-231 to human PrPSc.

(A) Capture of PrPSc or PrPC by rHuPrP23-231 was performed by incubation of rHuPrP23-231-conjugated magnetic beads with uninfected (CTL) and iCJD human brain homogenates, respectively. Magnetic beads without conjugated proteins were used as negative control (Empty). The g5p-conjugated beads were used as a positive control (g5p). To determine the specificity of the binding, we also examined the beads conjugated with recombinant PDI. The PK-resistant PrPSc was only detected in the preparations captured by rHuPrP23-231 and g5p beads from CJD brain homogenates. In contrast, no PK-resistant PrP was detected in CJD samples captured by empty beads and PDI beads. No PrP signal was detected in the uninfected samples captured by all the beads except the rHuPrP beads. The bands detected in the preparation captured by rHuPrP beads are expected to be the recombinant PrP itself. (B) To determine whether PrP detected in the preparation captured by rHuPrP from the uninfected brain homogenate shown in panel (A) was brain PrPC or rHuPrP itself, the capture experiment was also performed in the binding buffer alone without uninfected brain homogenate. The virtual same PrP bands were detected in both capture experiments in the absence and presence of uninfected human brain homogenates, suggesting that the detected PrP bands were from rHuPrP itself and no brain PrPC was captured. The blots were probed with 3F4. The results shown in (A) and (B) are a representative of two experiments.

Full size image

Discussion

The in vitro and in vivo conversion efficiency of PrPC into PrPSc can be significantly affected by the presence of additional PrP molecules that differ from the endogenous PrPC by as little as one residue6,28,11,12. The present study now demonstrates that a PrP molecule that shares the identical amino acid sequence with the PrPC substrate and PrPSc template also causes interference. It is worth noting that although the amino acid sequence is identical, recombinant PrP does not contain N-linked glycans or a GPI anchor. Our new results suggest that in addition to the amino acid sequence, glycosylation and the GPI anchor are important in mediating the conversion of PrPC into PrPSc. The unglycosylated and anchorless recombinant PrP appears to act as an inhibitor of the conversion process by preferentially binding to PrPSc.

PrPC is a glycoprotein with two non-obligatory, N-linked glycosylation sites at residues 181 and 197 and a GPI anchor24,29,30. Binding of heterologous PrPC to PrPSc can be influenced by PrPC glycosylation in a species-specific manner31,32. Moreover, using PMCA and a scrapie cell assay, Nishina et al reported that the stoichiometry of host PrPC glycoforms modulates the efficiency of PrPSc formation in vitro17. Specifically, their study demonstrated that while unglycosylated PrPC is required to propagate mouse RML prions, in a similar reaction, amplification of hamster Sc237 prions is inhibited by substoichiometric levels of homologous unglycosylated PrPC. This study provides direct in vitro evidence that changes in the PrP glycoform ratios can affect the efficiency of PrPSc formation in a species-specific manner. Recently, we observed glycoform-selective prion formation in unique sporadic and inherited forms of Creutzfeldt-Jakob disease (CJD) including variably protease-sensitive prionopathy (VPSPr) and familial CJD linked to a valine to isoleucine mutation at residue 180 (fCJDV180I)16. Although all four glycoforms are present, including di-, monoglycosylated at residue 181 (mono-181), monoglycosylated at residue 197 (mono-197) and unglycosylated PrP forms in the brain of VPSPr and fCJDV180I, only the mono-197 and unglycosylated PrP species were converted into PrPSc. The mono-181 and diglycosylated PrP species were not converted into PrPSc in the cerebral cortical brain areas examined. Moreover, the level of the classic PK-resistant PrPSc probed with the 3F4 antibody was significantly decreased compared to typical sporadic CJD. Instead, a unique five-step ladder-like electrophoretic profile of PK-resistant PrPSc was detected in both diseases by the 1E4 antibody16. In contrast to the threonine to alanine mutation at residue 183 of PrP (PrPT183A), the PrPV180I mutation exhibits a typical PrP glycosylation profile, although there is no detectable mono-181 and diglycosylated PrPSc33,16. However, using the N-linked glycosylation prediction algorithm NetNGlyc 1.0 at http://www.cbs.dtu.dk/services/NetNGlyc/34, we predicted a slight decrease in the glycosylation potential at N181 in PrPV180I compared to PrPWt (0.597 vs 0.664) while no potential change was predicted at all for N181 in PrPT183A16. The prediction data suggests that although the T183A mutation completely eliminates the N181 glycosylation site, the V180I mutation may merely alter the glycan composition at N181, which modifies the ratio of the four PrP glycoforms in the PrP mixture. Further investigation into the mechanism, by which altered glycosylation affects both conversion efficiency of PrPC into PrPSc and PrPSc conformation, is warranted. Unglycosylated and anchorless recombinant human PrP may have greater affinity for PrPSc seeds compared to the brain PrPC, but it is a poor substrate for conversion into PrPres by the standard PMCA protocol or in ScN2a cells. Indeed, both recombinant hamster and mouse PrP are not converted into PK-resistant PrP by serial PMCA in the presence of hamster prion Sc237 or mouse prion RML, respectively17. It is worth noting, however, that the recombinant PrP could be converted into PrPres using a modified PMCA protocol in which the conversion buffer contained 0.1% SDS and the normal brain-derived PrPC was replaced by recombinant hamster PrP as a substrate35,36 and the product of this reaction was proved to be infectious in animal bioassays37. Furthermore, in the absence of brain homogenates, recombinant PrP was converted by PMCA to highly infectious prions in the presence of additional cofactors such as phosphatidylglycerol and RNA38 or phosphatidylethanolamine39. Prion infectivity was also produced in Syrian hamsters by inoculating full-length recombinant hamster PrP that was converted into a cross-β-sheet amyloid conformation and subjected to an annealing procedure40.

We cannot rule out the possibility that the inhibition of human PrPSc amplification by recombinant human PrP results from the lack of the GPI anchor, although the GPI anchor of PrPC is believed to have little or no effect on the formation of PK-resistant PrP31,32. Anchorless PrP generated in either cultured mammalian cells or E. coli are converted to PK-resistant PrP by a cell-free conversion approach41,42,43,44. Moreover, it has been reported that anchorless prion protein induced an infectious amyloid disease in transgenic animals, although the animal themselves were asymptomatic45. However, amplification of hamster PrPSc using a standard PMCA protocol is inhibited when the substrate of normal hamster brain PrPC was pretreated with phosphatidylinositol-specific phospholipase C (PIPLC) to remove the GPI anchor19. Furthermore, recombinant hamster PrP was previously shown to inhibit PMCA of hamster PrPSc using the normal hamster brain homogenate as a substrate18. Kim et al proposed that both of these effects are due to the lack of the GPI anchor in PIPLC-treated PrPC or recombinant PrP19. On the other hand, the unglycosylated hamster PrPC purified from brains and containing an intact GPI anchor likewise inhibits amplification of Sc237 prions17. Therefore, the role of GPI anchor in the inhibition of human and mouse PrPSc propagation by recombinant human PrP observed in our study remains to be determined.

We demonstrated that recombinant human and other PrP that exhibited 50% or greater inhibition of PrPSc formation in a PMCA reaction bind to human PrPSc but not to PrPC. Although full-length, N- or C-terminally truncated recombinant PrP all bind to PrPSc efficiently, the full-length rHuPrP23-231 exhibits the highest inhibition efficiency compared to the two truncated forms, suggesting that the inhibition involves both N- and C-terminal domains. Moreover, antibodies including SAF32, 3F4 and 6H4 directed against PrP regions covering residues 59 to 152 showed less than 10% inhibition. The 8H4 antibody against human PrP175-185 exhibited virtually no inhibition. These results are in good agreement with a previous report by Horiuchi and Caughey1. In addition, it has been shown that the 3F4 and 6H4 antibodies preferentially bind to native PrPC, although they also detect denatured PrPSc on Western blots46,47. Therefore, the interaction of inhibitors with PrPSc may be required for the inhibition of PrPC conversion. We observed that recombinant mouse PrP is also able to bind to human PrPSc (Figure S2), although it caused significantly less inhibition compared to recombinant human PrP. Moreover, the anti-DNA antibody that specifically captures PrPSc but not PrPC showed less than 10% inhibition while g5p caused more than 50% inhibition. This suggests that the different inhibitors have distinct binding sites on PrPSc: one class of sites is specifically associated with recruiting PrPC while the other is not. Recombinant human PrP is likely to compete with brain PrPC for the same site on the PrPSc molecule. Moreover, its affinity for PrPSc seems to be greater than that of brain-derived PrPC. Interestingly, a two-site model has been proposed by Horiuchi and co-workers to explain the molecular mechanism for sequence-difference interference28. According to this model, PrPSc has two types of PrPC binding sites: one is able to induce conversion to PrPSc while the other is not.

A recombinant mouse PrP with a substitution of lysine for glutamine at mouse codon 218 (rPrP-Q218K), corresponding to human PrPE219K, an Asian-specific polymorphism believed to be resistant to CJD infection, considerably prolonged incubation time of prion infection in an iatrogenic mouse model48. Recombinant mouse PrP was delivered into the mouse brain for 7 days by intracerebroventricular administration using an indwelling catheter connected to an implanted osmotic pump. The same group also found that rPrP-Q218K reduced PrPSc formation in ScN2a cells. However, using wild-type mouse PrP did not cause inhibition, which is different than our findings. This discrepancy may be due to different experimental conditions between the studies. We showed that murine PrPSc amplification was inhibited in both PMCA and ScN2a by unglycosylated and anchorless recombinant human PrP. Most importantly, since the amino acid sequence of recombinant human PrP is identical to that of human brain PrPC, it is expected that this protein would not elicit an immune response after intracerebroventricular administration while it inhibits PrPSc propagation. Therefore, our findings suggest a new therapeutic strategy for treating human prion diseases.

Methods

Recombinant prion protein, protein disulfide isomerase and mechlorethamine

The various constructs for producing recombinant protein: human PrP [rHuPrP23-231 or rHuPrP90-231 with methionine at polymorphic residue 129 (129M)], mouse PrP (rMoPrP23-231), or bank vole PrP [rBvPrP23-231 with isoleucine at polymorphic residue (109I)] were cloned into the pET28a vector (Merck Millipore) and expressed and purified as a soluble protein as previously described49. Another set of recombinant human PrPs including rHuPrP23-231 and rHuPrP90-231 with 129M or 129V described previously23 was used to confirm the results with recombinant PrP generated in this study. The protein disulfide isomerase (PDI) plasmid was a generous gift from Dr. Joris Messens. Expression and purification of PDI were performed as previously described50. rHuPrP (23-145) was kindly provided by Dr. Giuseppe Legname. Recombinant mouse (rMoPrP23-231, the second source) and bovine PrP (rBoPrP23-231) were purchased from Prionics AG (Zurich, Switzerland). Mechlorethamine (MCT) was purchased from Sigma-Aldrich (Milwaukee, WI, USA).

Construction of transgenes expressing human PrP-129V

The transgene constructs were based on the murine half-genomic PrP clone in plasmid pHGPRP51. The HuPrP-129V open reading frame (ORF) was amplified from the human genomic DNAPAC (P1-derived artificial chromosome) clone RP5–1068H6 (obtained from the Sanger Center, Cambridge, UK) with primers HRM-F (TATGTGGACTGATGTCGGCCTCTGCAAGAAGCGC) and HRM-R (CCACCTCAATTGAAAGGGCTGCAGGTGGATAC). The PCR product was digested with PshAI and MfeI and used to replace the corresponding 0.97 kb PshAI–MfeI fragment in pHGPRP to create pHGHuPrP-129V. The inserted 0.97 kb PshAI–MfeI fragment in pHGHuPrP-129V was then sequenced with the primers HRM-R, HRM-F and HP306R (CATGTTGGTTTTTGGCTTACTC). One error free clone was chosen for the creation of transgenic mice.

Generation, screening and characterization of transgenic Tg(HuPrP-129V)Prnp0/0 mice (TgWV)

The 12.2 kb HuPrP-129V transgene construct was microinjected into fertilized FVB/NJ eggs and planted into the oviducts of pseudopregnant CD-1 mice at the transgenic mouse facility of Case Western Reserve University (Cleveland, OH). Founder pups were screened by PCR of tail DNA. All founder mice carrying the transgene were bred with FVB/Prnp0/0 mice51 to obtain Tg mice in PrP-null background. Transgenic PrP expression in the brain and other tissues of the Tg mice was examined by Western blot analysis using monoclonal antibody 3F4 for humanized Tg mice. All animal experiments in this study were approved by the Institutional Animal Use and Care Committee and the Institutional Biosafety Committee.

Preparation of brain homogenates

Frozen brain tissues from the frontal cortex of a patient with iatrogenic CJD (iCJD) and uninfected normal controls were obtained at autopsy. The characterization of this iCJD case has recently been reported52. Consent to use the autopsy brain tissue had been obtained in advance. The use of human brain tissues was authorized by the Institutional Review Board. Brain tissues from a mouse infected with mouse prion strain 139A and FVB wild-type mice were also used. A10% (w/v) infected human or mouse brain homogenate was prepared as described previously21,53, which was used as the PrPSc template for PMCA. To prepare the substrate for PMCA, brains from humanized transgenic mice described above were perfused with 5 mM EDTA in PBS and a 10% (w/v) brain homogenate was prepared as described21.

Protein misfolding cyclic amplification

PMCA was performed as described with slight modifications54,21. In brief, the samples were subjected to PMCA, consisting of cycles of 30 min incubation at 37°C followed by a 40-second pulse of sonication at 60% potency for 18 h in a sonicator (QSONICA 700, Newtown, CT). To detect the amplified PrPSc, 20 μl of PMCA-treated or untreated sample was incubated with 100 μg/ml PK for 70 min at 45°C. The reaction was terminated by adding PMSF to a final concentration of 5 mM and an equal amount of SDS sample buffer. Samples were then heated at 100°C for 10 min and a 10-μl sample was subjected to SDS-PAGE and Western blotting with 3F4 or 6D11.

Scrapie-infected mouse neuroblastoma cell culture (ScN2a)

The effect of rHuPrP on PrPSc propagation in ScN2a cells was conducted as previously described with a minor modification25,26. Briefly, ScN2a cells seeded in six-well plates (5 × 105 cells/well) containing 3 mL supplemented Opti-medium + 5%FBS were incubated with designated concentrations of recombinant proteins for 4 days. The cell lysates were prepared as previously described34. Samples of equal volumes containing equivalent amounts of protein were digested with 25 μg/mL proteinase K for 1 h at 37°C. Digestion was stopped by addition of PMSF to a final concentration of 2 μM and an equal volume of sample buffer was added. The samples were boiled for 10 min before loading onto 15% SDS-PAGE (SDS-polyacrylamide gel electrophoresis) precast Criterion gels (Bio-Rad, Hercules, CA, USA).

Specific capture of abnormal PrP by g5p, recombinant PrP or anti-PrP antibodies

The preparation of g5p-, rPrP-, rPDI-, anti-PrP antibodies-magnetic beads and capture of PrPC and PrPSc by the conjugated beads were performed as previously described22,27.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting

SDS-PAGE and immunoblotting was conducted as previously described27.

Statistical analysis

Statistical significance of differences in PrP intensity was evaluated using Student's t-test. A difference was considered statistically significant if the p value was <0.05.