The anti-prion activity of Congo red. Putative mechanism.

PrPSc, an abnormal conformational isoform of the normal prion protein, PrPC, is the only known component of the prion, a proteinacious agent that causes fatal neurodegenerative disorders in humans and other animals. The hallmark properties of PrPSc are its insolubility in nondenaturing detergents and its resistance to digestion by proteases. Anions such as Congo red (CR) have been shown to reduce the accumulation of PrPSc in a neuroblastoma cell line permanently infected with prions as well as to delay disease onset in rodents when administrated prophylactically. The mechanism by which such anti-prion agents operate is unknown. We show here that in vitro incubation with CR renders native PrPSc resistant to denaturation by boiling SDS. This resulted from PrPSc conformation, since neither the properties of PrPC nor those of predenatured PrPSc were changed by the addition of CR. CR-PrPSc could only be denatured by the addition of acidic 3 M guanidine thiocyanate. Since in vitro conversion experiments have suggested that partial denaturation may be required for PrPSc to serve as template in the PrPC --> PrPSc conversion, we propose that CR inhibits prion propagation by overstabilizing the conformation of PrPSc molecules.

Prion diseases, such as Creutzfeldt-Jakob disease in humans or scrapie and bovine spongiform encephalopathies in animals, are fatal transmissible neurodegenerative diseases (1). The only component of the prion is PrP Sc , an abnormal conformer of the host encoded PrP C . Both PrP isoforms share the same amino acid sequence, but, as opposed to PrP C , PrP Sc is relatively resistant to digestion by proteases and insoluble in nondenaturing detergents (2). Unlike PrP C , which is mainly ␣-helical, PrP Sc contains a considerable amount of ␤-sheet structures (3)(4)(5). Both isoforms are localized in cholesterol-rich membrane microdomains called rafts or caveolaelike domains (6 -8). The cell biology of prion diseases as well as the metabolism of the prion protein isoforms, have been investigated extensively in scrapie-infected mouse neuroblastoma cells (ScN2a) (9,10). These cells are permanently infected with prions and produce both PrP Sc and PrP C . In addition, inoculation of ScN2a extracts into mice causes prion disease (11). It has been postulated that prion diseases propagate by the con-version of ␣-helical PrP C molecules into ␤-sheet enriched PrP Sc by a mechanism in which PrP Sc serves as a template (12)(13)(14). Such a pathway may feature either the formation of PrP C -PrP Sc heterodimers alone or be also aided by mediator molecules. Experiments with transgenic mice favor the presence of such mediators and predict some of them to be species-specific (15).
While there is currently no effective therapy for prion diseases, an impressive body of data indicates that polyanions in general, and sulfated sugars in particular, inhibit prion propagation both in animals and in cultured cells. An array of such reagents have been utilized in an effort to "cure" prion infection, either in ScN2a cells or in animal models (16 -19). The criteria for the "cure" of ScN2a cells are reduction of cell infectivity titers and decrease of PrP Sc levels produced by the cells. Several anti-prion agents, such as Congo red (CR), 1 pentosan polysulfate, and others, fulfill these criteria (20). In rodents, coinoculation of prion samples with CR (21), amphotericin B (22), and lately iododoxorubicin (19) resulted in prolonged incubation time, a phenomenon that represents reduction in effective prion titers.
The mechanism by which these anti-prion agents operate has not been elucidated, although it can be assumed that they must interfere with one or more of the steps leading to the completion of the PrP C 3 PrP Sc conversion. Diverse sulfated polyanions seem to interfere with the subcellular metabolism of prions in the host cell, for example by altering the recycling of PrP to the interior of the cell (20). Interestingly, none of these reagents were shown to inhibit PrP C production (23). Some of them, such as CR, pentosan polysulfate, or low molecular weight heparin, were shown to inhibit the binding of PrP to heparin (17,24). Interestingly, negatively charged molecules and, in particular, heparan sulfate are found in plaques of every amyloid protein composition, and it has been proposed that cellular heparan sulfate proteoglycans may be involved in the replication of prions (25). The fluorescent antracycline iododoxorubicin was also shown to decorate amyloid deposits in brain sections of Creutzfeldt-Jakob disease patients (19). Thus, it is possible that some of these reagents interact directly with the prion proteins in the inoculum.
In this work, we investigated whether CR inhibition of prion propagation may result from direct interaction of CR with PrP Sc . The activity of CR as an anti-prion agent is particularly interesting, since this reagent is used as an amyloid dye in many systems including prion diseases (26). In fact, its binding to protein aggregates is considered diagnostic of an amyloido-genic process (27). Our results show that such an interaction of CR with the prion isoform exists and results in overstabilization of the PrP Sc molecules, as can be seen by their increased resistance to denaturation. If so, CR inhibition of prion propagation may result form the inability of CR-PrP Sc to convert into the partially denatured molecules required as template in the PrP C 3 PrP Sc reaction.

EXPERIMENTAL PROCEDURES
Chemicals and Tissue Culture Reagents-Congo red (98%) purchased from Aldrich was dissolved in distilled water and sterilized before use. Reagents for cell culture were from Biological Industries (Beit Haemek, Israel). G418 was from Calbiochem. ␣-L-Phosphatidylcholine was obtained from Avanti Biochemicals. Secondary antibodies were either from Promega (Madison, WI) or Jackson Immunoresearch (West Grove, PA). All other reagents were from Sigma.
ScN2a-c10 Cells Cultured with CR and Tunicamycin-Identically seeded and confluent ScN2a-c10 cells were cultured in 10-cm plates (about 10 7 cells) and incubated for 16, 24, or 48 h in the presence of either 50 g/ml CR, 1.5 g/ml tunicamycin, or both. Cells were extracted in 1 ml of lysis buffer containing 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, and 1% Nonidet P-40. The samples were centrifuged at 3000 rpm for 15 min at 4°C. The supernatant was concentrated by methanol precipitation and digested with 40 g/ml proteinase K (PK) for 20 min at 37°C. 3 M acidic GndSCN was added to half of each sample before boiling in SDS sample buffer and immunoblotting with anti-PrP Ab 3F4.
Incubation of Syrian Hamster Brain Membranes with CR-Normal and scrapie-infected Syrian hamster brains were kindly provided by Dr. S. B. Prusiner from the University of California, San Francisco. Membranes were prepared as follows. One brain (1 g) was homogenized in 10 ml of cold sucrose buffer (0.3 M sucrose, 10 mM Tris-HCl, pH 7.5) on ice. The homogenate was centrifuged at 2000 rpm for 15 min, and the supernatant was subsequently supplemented with 40 ml of sucrose buffer and centrifuged at 20,000 rpm for 30 min at 4°C. The pellet was resuspended in with 40 ml of cold 10 mM Tris-HCl, pH 7.5, and centrifuged again at 20,000 rpm for 30 min. The final pellet was resuspended to a concentration of 5 mg/ml in STES buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 2% sarkosyl).
Normal and scrapie-infected hamster brain membranes were incubated with increasing concentrations of CR for 1 h at room temperature. Scrapie samples were digested with 40 g/ml PK for 20 min at 37°C, and all samples were boiled in SDS sample buffer before immunoblotting with mAb 3F4. The percentage of acrylamide for SDS-PAGE (10% for the scrapie samples or 12% for the normal samples) was determined to avoid comigration of the PrP band with the dye.
Incubation of Prion Rods with CR-Syrian hamster prion rods (PrP- (27)(28)(29)(30)) purified by the sucrose gradient method (32) were kindly provided by Dr. S. B. Prusiner. PrP- (27)(28)(29)(30) was diluted to a final concentration of approximately 100 ng/ml in STES buffer and incubated with increasing concentrations of CR for 1 h at room temperature. Half of each sample was then incubated with 3 M guanidine thiocyanate, pH 2.5, for 10 min and concentrated by methanol precipitation prior to SDS-PAGE analysis and immunoblotting with mAb 3F4.
PK Digestion of Denatured PrP- (27)(28)(29)(30)-Purified PrP- (27)(28)(29)(30) was incubated in the presence or absence of 200 g/ml CR for 1 h at room temperature. After the incubation, the samples were boiled in 1% SDS and then diluted 10-fold with STES buffer. Samples were digested with 20 g/ml PK for 30 min at 37°C and treated with or without 3 M acidic GndSCN as described above.
Immunoprecipitation of PrP- (27)(28)(29)(30) in DLPC-150 ng/ml of purified PrP- (27)(28)(29)(30) in DLPC were incubated in the absence or presence of 200 g/ml CR for 1 h at room temperature and immunoprecipitated with mAb 3F4 (diluted 1:250) for 18 h at 4°C. Protein A-Sepharose was then added, and the mixture was incubated for an additional 30 min at room temperature. The beads were rinsed five times in TNS (100 mM NaCl, 1% sarkosyl, 10 mM Tris-HCl, pH 7.5) and were separated from the supernatant. Supernatant and beads were treated in the presence or absence of 3 M acidic GndSCN followed by methanol precipitation. All samples were boiled in sample buffer, which did not contain ␤-mercaptoethanol, and immunoblotted with antiserum RO73.

CR Renders PrP Sc Resistant to Denaturation by Boiling
SDS-To investigate whether direct interaction of PrP Sc with CR may result in inhibition of prion propagation, we tested the biochemical properties of the PrP proteins after incubation of membranes from normal or scrapie-infected hamster brains with CR. The hallmark property of PrP Sc is its protease resistance. While PrP C is highly sensitive to proteases, digestion of PrP Sc with PK results in PrP- (27)(28)(29)(30), the protease-resistant core of PrP Sc (34). To test whether incubation of PrP Sc with CR will render this protein sensitive to PK, normal and scrapie brain samples were incubated with increasing concentrations of CR. Scrapie-infected hamster samples presumably contain both PrP isoforms, as opposed to normal brain membranes that contained only PrP C (35). To differentiate the protease-sensitive PrP C from protease-resistant PrP Sc , the scrapie-infected brain samples (CR-treated and controls) were digested with PK (see "Experimental Procedures") prior to PAGE analysis. All samples were subsequently boiled in SDS sample buffer and processed for immunoblotting. While the PrP C pattern on the gel carrying only normal brain samples was not affected by the preincubation with CR (Fig. 1A), the presence of PrP- (27)(28)(29)(30) on the immunoblot loaded with samples from scrapie-infected brains was reduced in a dose-dependent manner (Fig. 1B). These results suggested the intriguing possibility that direct incubation of the scrapie isoform with CR in vitro is enough to render PrP Sc sensitive to digestion by proteases.
This attractive possibility, however, was discarded after the same experiments were repeated with purified PrP- (27)(28)(29)(30). Purified PrP-(27-30) was incubated with CR (as described under "Experimental Procedures") and subsequently boiled in SDS sample buffer before immunoblotting ( Fig. 2A). As mem-FIG. 1. In vitro preincubation with CR differentiates PrP C from PrP Sc . Membranes from normal (A) and scrapie-infected (B) hamster brains (35) were diluted to 5 mg/ml protein with STES buffer and incubated with increasing concentrations of CR for 1 h at room temperature. After the incubation, scrapie samples were digested with 40 g/ml PK for 20 min at 37°C. Samples were boiled in SDS sample buffer before immunoblotting with mAb 3F4 (29). The percentage of acrylamide for SDS page (10% for the scrapie samples or 12% for the normal samples) was determined to avoid comigration of the PrP band with the dye.
branal PrP Sc in Fig. 1, PrP- (27)(28)(29)(30) was not apparent in the lanes loaded with CR-treated samples. The absence of CRtreated PrP-(27-30) from the immunoblot was obviously not a result of acquired protease sensitivity of the scrapie protein, since in this case, the samples were digested with PK during the PrP purification protocol (32) and not after the incubation in the presence of CR. Nor were these results due to loss of PrP immunoreactivity, since CR-treated PrP-(27-30) could not be visualized even when the samples were applied to a silverstained SDS gel (Fig. 3C). The effect of CR on the scrapie PrP isoforms most probably resulted from the specific scrapie conformation of the PrP protein. This could be concluded from the fact that PrP- (27)(28)(29)(30) could be visualized by immunoblotting if it was denatured either by boiling in SDS (Fig. 3B) or by the addition of 3 M guanidine thiocyanate (GdnSCN) (Fig. 3A), prior to but not after the addition of CR.
CR-treated PrP-(27-30) rods ( Fig. 2A, ϩ) and membranal CR-treated PrP Sc (Fig. 2C) could be denatured and subsequently visualized on an immunoblot, albeit only when 3 M acidic GndSCN was added to the samples prior to their boiling in SDS sample buffer. These results suggest that, as opposed to our first erroneous conclusion that implied CR renders PrP Sc protease-sensitive, in fact, CR-treated PrP Sc as well as CRtreated PrP- (27)(28)(29)(30) are substantially more difficult to denature than control PrP Sc samples. This increased resistance to denaturation hinders the entrance of the scrapie protein into the gel, even after boiling in the presence of SDS.
CR Does Not Interfere with the Dispersion of PrP- (27)(28)(29)(30) into DLPC-In addition to protease resistance, another hallmark property of the scrapie prion protein is its insolubility in nondenaturing detergents (Fig. 4A). The insolubility of PrP Sc has been utilized in the protocol developed for its purification as PrP-(27-30) rods (37). Interestingly, as opposed to its behavior in many detergents, the insoluble rods could be dispersed into DLPC (33). When PrP- (27)(28)(29)(30) in prion rods was sonicated in the presence of phosphatidylcholine and 2% cholate and subsequently centrifuged at high speed, most of the PrP remained in the supernatant, as opposed to the samples without lipids, which are pelleted at these conditions (Fig. 4B). Although it is difficult to conclude whether PrP-(27-30) is completely solubilized to monomers by this procedure, it is obviously more dispersed than in rods and, as a result of this, more infectious (38). To investigate whether CR competes with lipids for binding sites on PrP Sc , we tested whether CR-PrP- (27)(28)(29)(30) could also be incorporated into DLPC. As can be seen in Fig. 4B, when CR-treated PrP- (27)(28)(29)(30) was dispersed into DLPC by the same protocol used for control PrP, most of the CR-treated protein remained in the supernatant. As in the previous experiments, this could be visualized only after the addition of acidic Gnd-SCN to the samples prior to their analysis by PAGE (Fig. 4B). Moreover, PrP-(27-30) in DLPC was still susceptible to CR, since acidic GdnSCN was required to visualized on the immunoblot PrP-(27-30) to which CR was applied after the disper-  3 and 4) of 200 g/ml CR. After the incubation, all samples were digested with 20 g/ml PK for 30 min at 37°C. Prior to the PK digestion, samples 2 and 4 were boiled for 10 min in the presence of 1% SDS and then diluted 10-fold with STES buffer. Samples in the lower part were treated with 3 M acidic GndSCN and concentrated by methanol precipitation prior to SDS-PAGE. All samples were immunoblotted with anti-PrP mAb 3F4. C, membranes from scrapie-infected brains were treated as in Fig. 1B (500 g/ml CR). After PK digestion, half of each sample was incubated for 2 min with 3 M GndSCN, pH 2-2.5 (lower part), before SDS boiling.  (27)(28)(29)(30) was incubated in the presence or absence of 3 M GdnSCN followed by methanol precipitation. The precipitated samples were treated with and without CR as described in Fig. 2B (500 g/ml CR) and subsequently immunoblotted with mAb 3F4. B, PrP- (27)(28)(29)(30) was boiled in the presence of SDS for 5 min and subsequently diluted before incubation in the presence of 0, 100, or 500 g/ml CR. All samples were immunoblotted with mAb 3F4. C, PrP- (27)(28)(29)(30) was incubated in the presence and absence of 200 g/ml CR and applied to SDS-PAGE after boiling in SDS sample buffer. sion into DLPC (Fig. 4C). This result suggests that dispersion into DLPC did not hide the CR binding site of PrP- (27)(28)(29)(30).
To test whether the CR addition to PrP in DLPC results in reaggregation of the protein, we recentrifuged the samples after the addition of CR to previously dispersed PrP-(27-30) (Fig. 4D). PrP- (27)(28)(29)(30) remained in the supernatant after an additional high speed centrifugation. Again, the prion protein could be visualized only after the addition of acidic GdnSCN to the sample buffer.
Dispersion of native PrP Sc or PrP- (27)(28)(29)(30) into DLPC significantly increased their otherwise very poor affinity for PrP antibodies (28,39). To test the effect of CR on this antibody affinity, PrP- (27)(28)(29)(30) in DLPC as well as CR-treated PrP- (27)(28)(29)(30) in DLPC were incubated with either mAb 3F4 or mAb 13A5 (29,31) followed by precipitation of the immunocomplexes with protein A-Sepharose (Fig. 5). Acidic GndSCN was added to the samples prior to boiling in SDS sample. All samples were thereafter immunoblotted with polyclonal PrP antiserum RO73 (28). While a significant fraction of PrP- (27)(28)(29)(30) in DLPC was immunoprecipitated by either of the ␣-PrP mAbs, almost no PrP was present in the pellet of the CR-PrP sample. These results are consistent with the possibility that CR may stabilize the PrP Sc configuration in such a way that its binding to other cell components, or even to its own antibodies, is now inhibited.
CR Inhibited New PrP Sc Synthesis and Old PrP Sc Degradation-The best cell culture model developed to date for prion infection is the ScN2a system. These scrapie-infected mouse neuroblastoma cells produce PrP Sc as well as infectivity (11,40). When ScN2a cells were cultured in the presence of CR, no new PrP Sc was synthesized and subsequently accumulated (41). We tested the fate of PrP Sc in these cells to resolve whether in previous experiments some of the "old" scrapie protein may not have been accounted for because of the insufficient denaturation that results from the addition of CR. ScN2a-c10 cells, a ScN2a clone expressing a chimeric mouse/ hamster PrP (47), were incubated with and without CR and in the presence or absence of tunicamycin, a glycosylation inhibitor, for 18, 30, or 48 h. Cell lysates were then digested with 20 g/ml PK for 30 min at 37°C. In the presence of tunicamycin, new PrP is synthesized as an unglycosylated protease-resistant protein (42), while fully glycosylated PrP Sc , higher in M r , present prior to the addition of tunicamycin is degraded very slowly. Such an experiment is shown in Fig. 6. In the absence of CR, fully glycosylated "old" PrP (upper band of the three that comprise the PrP immunoblot profile) was partially degraded after 48 h in tunicamycin, while the lower band signal (representing unglycosylated PrP) was considerably enhanced as compared with the signal obtained before the addition of tunicamycin. When tunicamycin was added to the cells incubated in the presence of CR, no new protease-resistant PrP was synthesized, as seen by the fact that there was no increase in the signal of the low molecular weight band. In effect, no difference at all could be appreciated in the banding pattern of PK-resistant PrP in the presence or absence of tunicamycin. These results suggest that no new PrP Sc was made by ScN2a-c10 cells in the presence of CR, although the concentration of the old PrP Sc remained high. As with the previous experiments performed on membranes or purified PrP- (27)(28)(29)(30), the total concentration of PrP Sc in the CR-treated cells could be visualized only when acidic GndSCN was added to the denaturing sample buffer.

DISCUSSION
Prions appear to propagate by refolding mature PrP C , a membrane glycoprotein of unknown function, into an isoform of aberrant conformation denominated PrP Sc . This refolded protein is the only known component of the prion. Thus, the two isoforms of PrP serve, at the same time, as substrate and template in the conversion process that results in the accumulation of more prions. During the conversion, PrP Sc somehow confers to PrP C its own and distinct structural conformation, which is extremely stable and comprises large ␤-sheet segments (5). Although the very nature of the conversion process is unknown, in vitro studies have suggested that partial denaturation of PrP Sc may be a crucial requirement for this molecule to act as a template in the further conformational conversion of PrP C (43). In addition, theoretical considerations also suggested the involvement of a partially denatured molten globule in the formation of PrP Sc (44). The requirement of the chaperon HSP104 for the propagation of the prion like ⌿ of yeast also suggested that template "breathing" may feature in the propagation of conformers (45). Therefore, it is possible that reagents that stabilize PrP Sc so that it cannot be partially denatured will inhibit prion propagation. Our results show that CR can act directly on PrP Sc , even in vitro, and that its effect can be evaluated in a simple assay. Since reagents such as CR inhibit further propagation of prions, they must interfere with one or more steps in the conversion process. While other anti-prion agents may disrupt the PrP Sc conformation and render this protein protease-sensitive, as expected from an anti-prion agent, CR seems to reinforce PrP Sc structure and, as a result of this, inhibit its denaturation. In view of this, we suggest that CR may hinder prion propagation and PrP Sc formation by stabilizing the template PrP Sc so that the partial denaturation needed for the conversion of additional PrP C molecules is prevented.
The effect of CR on PrP Sc is not a function of the PrP sequence or its glycosylation profile but rather of its conformation. We conclude this from the fact that neither PrP C (Fig. 1) nor predenatured PrP Sc (Fig. 3) changes its basic properties when incubated in the presence of CR. It has been shown in the past that CR does not inhibit the synthesis of PrP C (24).
CR association abolished the affinity of PrP- (27)(28)(29)(30) to its antibodies. This affinity, very poor in purified PrP rods, was significantly enhanced only after PrP-(27-30) dispersion into DLPC. In addition, the comparison of the lines in Fig. 6 with and without guanidine denaturation suggests that inhibition of synthesis of new PrP Sc by CR occurs even in the presence of the PrP Sc molecules that were present in the cells before the addition of CR.
Our results are consistent with yet another plausible conclusion. Since in the presence of CR, PrP Sc seems to be "invisible" to either antibodies or metabolic cell components, it may also be invisible to components of the PrP C 3 PrP Sc conversion site. Whether this effect is due to direct inhibition of binding by CR or is also a result of the increased stability conferred to PrP Sc by CR remains to be established.
Scrapie incubation time may be determined not only by the concentration of PrP Sc in the inoculum, but also by the number of PrP Sc molecules available for binding to conversion sites. Indeed, PrP- (27)(28)(29)(30) in DLPC, a more dispersed form of PrP, has been shown to be more infectious, by a factor of 10 or 100, than the same sample in the more aggregated form (33,46). Conversely, it is possible that highly aggregated as well as metabolically stable PrP will be less infectious, since it cannot interact efficiently with the crucial cell components needed for the conversion of PrP C . This is probably why it has been very difficult to correlate PrP Sc concentrations with infectivity titer (47). As stated above, ScN2a cells did not produce new PrP Sc in the presence of CR, although acidic guanidinium revealed the presence of large concentrations of very stable PrP Sc .
The results presented here suggest an interesting feature of transmissible amyloid diseases. While it is necessary for the aberrant isoform of the amyloidogenic protein to be stable enough to metabolic degradation, so that it can be present long enough in the cell to transform many normal molecules, an overstabilized aberrant conformation will inhibit the partial denaturation required for template formation. Interestingly, the amyloid aggregates from a nontransmissible brain neurodegenerative disorder, Alzheimer's disease, cannot be denatured by SDS boiling but by much harsher conditions such as 6 M urea (48).
Whether the anti-prion effect of CR is caused by stabilizing the scrapie conformation only or also by blocking the binding of PrP Sc molecules to conversion sites, the fact that direct interaction of CR with PrP Sc inhibits new PrP Sc synthesis reinforces the notion that PrP Sc is probably the crucial if not the only prion component.
What molecules in the cell, in addition to PrP C , participate in the conversion reaction? Several molecules have been shown to either bind to PrP or be altered in scrapie-infected cells or brains (49 -52). Experiments with transgenic mice have postulated the existence of a species-specific factor, denominated Protein X, which is yet to be identified (15). The presence of glycosaminoglycan molecules such as heparan sulfate in every amyloid was proposed as evidence of their possible role in the scrapification of PrP (25). However, the fact that PrP C , although at a low efficiency, can be converted in vitro to a PrP Sclike molecule (43), suggests that such molecules may be important mostly for the kinetics of the reaction or for the pathogenesis of the prion disease. As proposed above, it is the very nature of PrP Sc conformation that probably hinders the finding of such molecules.
Elucidating the diverse mechanisms by which the known anti-prion agents operate may help in understanding the biology of prion diseases and may also be a crucial step for the preparation of reagents that can be used in in vivo experiments and clinical trials. Some of them will eventually also be useful for the treatment of other amyloid disorders.