Some notes on Scrapie

Chris Seidel     September 1995

     Scrapie is a member of a class of neurodegenerative diseases known as spongiform encephalopathies. Scrapie is referred to as a disease of protein conformation because it appears to be caused by the misfolding of a normal cellular protein, termed the prion protein, or PrP (Prusiner, 1982). While scrapie was originally identified in sheep and goats, the prion diseases span from chronic wasting disease in deer and elk, bovine spongiform encephalopathy (BSE) of cattle, transmissible mink encephalopathy (TME), and several diseases in humans such as Kuru, Creutzfeldt-Jakob disease (CJD), Gerstmann- Straussler-Scheinker syndrome (GSS), and fatal familial insomnia (Gajdusek 1988). Common among these disorders is the three stage transition of PrP from normal cellular protein to the scrapie isoform and then to scrapie amyloid fibers which accumulate extracellularly in the brain (Gajdusek, 1988; Prusiner, 1992). The prion diseases have been something of an enigma due to their mysterious nature of transmission and propagation. Scrapie and Kuru are infectious and can be spread in a variety of ways, including cannibalism. There is also an associated incubation time between exposure and sickness that can range from months to decades.

     Because scrapie is infectious, early investigators were led to propose that scrapie was a viral agent with a slow course of infection (Cho, 1976; Kimberlin & Hunter, 1967). However the search for an associated nucleic acid has not been fruitful. The scrapie agent appears transparent to UV radiation at 254 nm and easily maintains infectivity at doses of UV far beyond those required to inactivate any virus tested (Alper et al.,1967). However, since different strains of Scrapie can be isolated, some investigators adhere to the notion of an associated independent genome for scrapie propagation (Bruce & Dickinson, 1987; Braig & Diringer, 1985).

     The scrapie agent was identified by enriching fractions from Syrian hamster brains for scrapie infectivity (Prusiner et al., 1982). The subsequent identification of a scrapie specific protein and generation of antibodies led to the identification of the prion protein in both infected and uninfected brains. This finding suggested for the first time that the scrapie agent is a single protein that exists as a different isoform in healthy versus uninfected tissue (Oesch et al., 1985). Subsequent studies have shown that the prion protein appears to be responsible for both the toxicity and infectivity of scrapie. The prion protein in its non-toxic isoform is referred to as PrPC, whereas the toxic and infective isoform is referred to as PrPSc. The two forms of PrP are easily distinguished based on their differential sensitivity to proteases, and solubility in detergents. PrP is normally observed as a 33-35 kDa protein in cellular extracts from either infected or healthy cells (Pfeifer et al., 1993). However PrPC is rapidly and completely degraded by digestion with proteinase K whereas PrPSc is stable to digestion with proteinase K, forming a 27-30 kDa protease resistant core termed PrP27-30. In addition, PrPC remains soluble in the presence of detergents, whereas PrP27-30 polymerizes into fibrils.

     The scrapie protein is encoded by a single copy cellular gene termed the prion protein gene (Oesch et al., 1985). The PrP gene is highly conserved across species, showing 88%-90% conservation at the genomic DNA level between sheep and other mammals such as human, rat, mouse, hamster and cow. The protein is encoded by a single exon, ruling out differential splicing as a cause of different protein conformers (Basler et al., 1986). The mRNA for PrP is expressed at similar levels in the brains and other tissues of both sick and healthy animals (Oesch et al., 1985). The normal cellular function of PrP is unknown. The protein is normally found as a membrane bound sialoglycoprotein (Stahl et al., 1987, Haraguchi et al., 1985) on the surface of neuroblastoma cells. Studies of mice in which the PrP gene has been removed (prn-P0/0) indicate that the gene is non-essential. Prn0/0 mice develop and behave normally with no visible or measurable defects (Bčeler et al., 1992). Curiously, loss of the PrP gene renders these mice no longer susceptible to scrapie infection (Bčeler et al., 1992). Furthermore, propagation of scrapie infectivity is also eliminated in mice missing the PrP gene (Prusiner, et al. 1993).

     In the past, studies of the scrapie reagent were greatly hindered by a bioassay requiring long incubation times on the order of months, or sometimes years, after injection of the agent into live animals, before a result could be seen (Prusiner 1992). With the development of chronically infected cell lines in which recombinant prion protein could be expressed and specifically identified (Butler et al., 1988), progress in studying scrapie has been greatly increased. Expression of prion protein in an infected cell line supports the conversion of endogenous PrPC as well as recombinant PrPC into PrPSc (Scott et al., 1992). While specific antibodies which can distinguish between the two isoforms have not been reported, the two forms are successfully distinguished based on their sensitivity to protease. Studies with cell lines have given insight into the cellular trafficking of the prion protein. However, the mechanism of conversion of prion protein into the scrapie isoform is not understood.

     Upon discovery that PrPC and PrPSc were encoded by the same gene a search began for some posttranslational modification that could give rise to the difference between the two proteins. However, examination of the two proteins for a covalent modification unique to one of the two isoforms has shown that both forms are modified identically. The prion protein undergoes several posttranslational modifications before either being degraded as PrPC or transforming into PrPSc and accumulating. A signal sequence present near the C-terminus targets prion protein synthesis to the endoplasmic reticulum (ER), where the sequence is cleaved leaving ser-231 as a site for attachment of a glycosyl-phosphatidylinositol (GPI) group. Addition of the GPI moiety serves to anchor PrP to the surface of the cell. The GPI anchor is present on both PrPC and PrP27-30. However, synthesis of a truncated PrP lacking the GPI anchor attachment signal still results in the production of PrPSc, thus eliminating this component as a necessary factor for PrPSc production (Rogers et al., 1993).

     PrP protein also undergoes glycosylation in the Golgi. There are two consensus sites for asparagine linked glycosylation near the C-terminus of the prion protein. Examination of the various forms of the prion protein reveals that PrPC, PrPSc, and PrP27-30 are all glycosylated. While a myriad of different oligosaccharide structures are possible on the prion protein (Endo et al., 1989), it is doubtful that they play any role in leading to the transformation of PrPC into PrPSc. Use of a scrapie- infected cell line which can support PrPSc production has shown that synthesis of PrPSc in the presence of agents which block N-linked glycosylation still allows the production of protease resistant PrPSc, albeit PrPSc produced in this fashion is unglycosylated (Taraboulos et al., 1990). These results have been repeated by other methods, including site directed mutagenesis to eliminate the sites of N-linked glycosylation. PrPSc is still produced in the absence of sites for N-linked glycosylation. Thus it would appear that glycosylation is not an essential element in the transformation of PrPC into PrPSc. To begin examining how early in the trafficking pathway scrapie could be detected, and assess whether events during protein processing were responsible for misfolding of PrP, the ER-Golgi was examined. Brefeldin A is a reagent which causes the cis-medial and trans-Golgi cisternae to fuse with the ER and prevents exit of proteins from the ER-Golgi network (Doms et al., 1989). Treatment of scrapie-infected cells with Brefeldin A to retain PrP within the ER-Golgi apparatus was found to reversibly block synthesis of PrPSc. Thus it would appear that compartmentalization of PrP within the ER-Golgi is insufficient for PrPSc formation. Removal of BFA allows the synthesis of PrPSc to continue.

     Shortly after exit from the Golgi, PrPC appears on the cell surface held to the plasma membrane via the GPI anchor. Treatment of cells with phosphatidylinositol-specific phospholipase C (PIPLC) cleaves the phosphatidyl linkage and releases protease sensitive PrPC into the medium. Interestingly this also inhibits production of PrPSc (Caughey & Raymond, 1991; Borchelt et al., 1990), suggesting that PrPSc is synthesized from PrPC after it travels to the plasma membrane. PrPC can be immunofluoresced on the surface of cells using an antibody that recognizes both PrPC and PrPSc. However upon treatment of cells with PIPLC, the immunofluorescence signal is lost (Borchelt et al., 1990). Although PrPSc is easily detected within cells (Taraboulos et al., 1990), it has not been detected on the surface of cells. Digestion of scrapie-infected cells with PIPLC does not release protease resistant PrP into the medium, yet the immunofluorescence signal is lost. Cell surface proteins can be labeled by a number of cell impermeant reagents (Goodloe- Holland & Luna ,1987). Exposure of scrapie-infected cells to sulfo-NHS-biotin allowed labeling of PrPC, but not PrPSc, although control experiments indicate that purified PrP27-30 is susceptible to sulfo-NHS- biotin labeling and maintains a protease resistant state afterwards (Borchelt et al., 1990). For these reasons it is thought that PrP enters the cell before conversion to PrPSc occurs.

     Evidence to date seems to indicate that the prion protein assumes the scrapie isoform in the endocytic pathway (Borchelt et al., 1992). However, direct identification of endosomes containing PrPSc has not been achieved. Immunofluorescence and immunocytochemical studies have been used to assay for the presence of PrP within cells. PrPSc is found to accumulate within cells, while PrPC is generally not observed except for very minor amounts in the ER-Golgi. From examination of the morphology of immunoperoxidase staining patterns, PrPSc was deduced to be associated with intracellular vesicles. No intracellular staining pattern was observed in healthy cells. These structures were found to be positive for acid phosphatase activity thus identifying them as secondary lysosomes. Kinetic studies show that the time between acquisition of protease resistance and exposure to lysosomal proteases is approximately one hour (Taraboulos et al., 1991). Use of lysosomotropic amines prevents the N-terminal trimming of PrPSc in the lysosome, yet does not block the formation of PrPSc, nor does it prevent the degradation of PrPC. Since PrPSc cannot be detected on the surface of cells, yet can be detected in lysosomes, and PrPC can not be detected in cells yet can be detected at the surface of cells, these observations taken together with those listed above point to the endosomal pathway as a possible place for the conversion of PrPC to PrPSc (Borchelt et al., 1992).

     In the absence of evidence for any significant covalent modifications that can distinguish PrPC from PrPSc, focus has shifted to the difference in secondary and tertiary structure between the two forms of the prion protein. Purification techniques for the scrapie protein have depended in large part on harsh conditions such as detergent extraction and proteinase K digestion. Purification of the prion proteins for structural studies would of course depend on more native conditions and structural studies have been hindered by the difficulty in purifying these two proteins. PrPC is soluble but difficult to purify in large amounts (Prusiner, 1991), and PrPSc is inherently insoluble and difficult to work with. Nonetheless spectroscopic data has been collected for both PrPC and PrPSc. Examination of PrPC and PrPSc by Fourier transform infrared (FTIR) spectroscopy shows that the two differ significantly in the relative amounts of alpha helix and Beta sheet that each species contains. These findings have been verified by Circular Dichroism spectroscopy as well. PrPC contains approximately 42% alpha helix, and 3% beta sheet. However, PrPSc contains a high amount of Beta sheet, approximately 43% which rises to 54 % in PrP27-30 the product of proteolytic processing of PrPSc (Pan et al., 1993).

     The mechanism of prion propagation is not understood. Several theories exist as to how the prion protein comes to be misfolded and form aggregates. There is general agreement that PrPC and PrPSc differ primarily by conformation. One model holds that scrapie formation is a kinetic event dependent primarily on seeded nucleation, much like protein crystallization (Come et al., 1993, Jarret & Lansbury, 1993), or the polymerization of hemoglobin S into fibers as in sickle cell anemia (Hofrichter et al., 1974). PrPC is postulated to unfold slightly, exposing a hydrophobic stretch of amino acids to solvent. Aggregation then depends on the colocalization of the seed and the monomer. The partially unfolded state, PrPU, is proposed to be easily accessible from PrPC. Since PrPU formation is considered rapid, PrPU monomer is readily available for aggregation in the presence of a seed. Formation of the seed is the rate limiting step, and is considered to be slow, and dependent on the concentration of monomer molecules. In the case of infection, seed is introduced from outside, after which rapid aggregation can begin. Sporadic disorders, on the other hand, would depend on seed formation from PrPU monomers. The equilibrium between PrPC and PrPU could easily depend on the cellular environment, and be influenced by things such as pH, or the presence of interacting factors as well as mutations in the PrP gene that favor PrPU formation (Brown et al., 1991).

     Other investigators hold to the notion that the prion protein can exist in a meta stable state capable of collapse to either of two more permanent states. This model put forth by Prusiner and colleagues suggests that PrPC can assume a partially unfolded intermediate, PrP*, which can then become PrPSc, be degraded, or resume the PrPC form. In addition, PrPSc would be able to serve as a template for the conversion of PrP* into PrPSc. This model is not dependent on aggregates or seeds for scrapie formation. The slow step is considered to be formation of PrP*. Under normal conditions the concentration of PrP* is low and formation of PrPSc is not significant. However, at some threshold level, there would be enough PrPSc present to form a complex with any PrP* that forms, thus converting it to PrPSc and catalytically increasing the concentration of PrPSc. Infection with PrPSc would allow propagation by increasing the concentration of PrPSc beyond the threshold level. Sporadic disorders would arise through events which cause an increase in the frequency of PrP*, and subsequently PrPSc, such as a mutation in the PrP gene, interaction with a co-factor, or an optimal cellular environment.

     Initial attempts to reconstitute the misfolding of PrPC into a protease resistant form in vitro were not successful (Raeber et al., 1992). Since it is postulated that the misfolded form serves as a template or agent to direct misfolding of the normal isoform, it is reasonable to suspect that these are the only two components required for the creation of protease resistant PrPSc. Mixing purified preparations of the two forms did not result in detection of a newly formed protease resistant species (Raeber et al., 1992). Mixing experiments involving crude extracts, or synthesis of PrP in the presence of microsomal fractions failed to generate PrPSc as well. However, recently investigators were successful at generating a protease resistant species of prion protein in vitro by mixing PrPC and PrPSc together. Furthermore, these results have recently been extended to include in vitro formation of different strains of scrapie-like proteins starting from a common PrP precursor. Both of these reactions were dependent on the use of guanidinium to unfold the template PrPSc protein slightly, but not enough to diminish its infectivity. Because these results were obtained using highly purified components, they offer very strong support of the protein-only model for scrapie propagation. The amounts of new scrapie produced are much too small however to test for infectivity expected with the acquisition of protease resistance. These results also lend support to the notion that different strains of scrapie may differ simply by conformation of the prion protein. The time course required for in vitro formation of protease resistant PrP, the inefficiency of the reaction, and the nature of having to slightly unfold the template molecule, if not also PrPC, suggest there are conditions in the cell that are optimal for this process, and may even include a co-factor such as a chaperonin. The role of chaperonins in directing the folding of proteins is becoming well documented though is not yet well understood.

     Recently, a dimer of PrP was isolated from mouse neuroblastoma cells expressing hamster PrP (Priola et al., 1995). The dimer molecule appeared to be covalently linked or was stable to several strongly denaturing reagents including boiling in SDS. Interestingly, the dimer molecule showed properties of both PrPC and PrPSc in that it was protease sensitive, yet also tended to form aggregates. Initial examination of the cellular trafficking of the PrP dimer indicates that it contains only high mannose glycans, suggesting that it does not traverse the Golgi where conversion to more complex glycans usually occurs. Furthermore, no PrP dimer could be detected at the plasma membrane of the cells. Thus it does not appear to follow the same trafficking pattern as normal PrP. It is suggested that the dimer molecule may be an intermediate in scrapie formation (Priola et al., 1995). According to the seeded nucleation model described above, the propensity of the dimer to aggregate offers a means to the formation of an ordered nucleus which could act as a seed, or be acted upon by a seed of PrPSc(Priola et al., 1995). Scrapie-infected hamster brains were found to contain the dimer in a protease resistant form. Furthermore, in a cell free system the protease sensitive dimer is capable being converted to the protease resistant form in the presence of PrPSc purified from scrapie-infected hamster brains (Kocisko et al., 1994). The observation that the PrP dimer does not appear to traverse or enter the Golgi is based solely on the finding that the dimer is sensitive to treatment with endoglycosidase H to assay the type of glycans it contains. Endoglycosidase H is able to cleave most high mannose glycans but is not able to cleave complex oligosaccharides (Tarentino et al., 1978). As proteins traverse the Golgi they undergo oligosaccharide processing. PrP is normally observed to become resistant to endoglycosidase H as it traverses the Golgi where its high mannose glycans are converted to complex glycans. While most proteins undergo processing to more complex glycans, the diversity of structures and the extent of glycan processing of proteins is variable. In addition, the conformation of a protein may influence whether it undergoes complex glycan processing or remains high-mannose. Thus it is possible that the dimer enters the Golgi and gets sorted to some pathway without undergoing the same glycosylation as PrP monomers. Although dimers of PrP have been reported in the past, in light of these recent results they are conspicuously absent from the majority of the literature on scrapie. However given the large amounts observed by Priola et al., and the initial characterization of the properties of the dimer, many questions remain to be answered about the possible role of the PrP dimer in scrapie formation, and a system appears within reach to answer these questions.

     Recent results examining the affect of scrapie on the distribution and expression of various heat shock proteins (Hsps) in scrapie-infected cells revealed effects on several of the Hsps in comparison to healthy cells. While the majority of chaperones are constituitively expressed and found in abundance in cells under normal conditions (Gething & Sambrook, 1992), some of the heat shock proteins are highly induced under cell stress. In mouse neuroblastoma cells two of the most highly induced proteins in response to stress are hsp28 and hsp72 (Tatzelt et al., 1995). Exposure of cells to a brief heatshock or to stress inducing agents such as sodium arsenite, or Azc, an amino acid analog of proline, results in a marked increase in expression of several major mammalian stress proteins in both healthy and scrapie-infected mouse neuroblastoma cells. However, scrapie- infected cells depart from the normal stress response in that they fail to induce hsp28 and hsp72. The reasons for lack of induction of these two proteins in scrapie-infected cells are not understood. Another member of the hsp70 family highly related to hsp72 is hsp73. Hsp73 is constituitively expressed and is usually found in the cytoplasm. Exposure of cells to stress results in the translocation of both hsp72 and hsp73 into the nucleus of the cell. In scrapie-infected cells, hsp73 shows an abnormal punctate staining pattern in the cytoplasm, reminiscent of the staining pattern exhibited by scrapie accumulation in cells (Tatzelt et al., 1995). In contrast to normal neuroblastoma cells, scrapie-infected cells exposed to stress inducing conditions fail to translocate hsp73 into the nucleus. How scrapie is able to compromise the cellular stress response by affecting these proteins is not understood. Tatzelt and colleagues speculate that hsp73 may become trapped in a subcellular compartment or form a tight complex with PrPSc. The hsp70 proteins are described as undergoing a host of interactions with proteins, binding both folded and unfolded proteins. These interactions may be important in scrapie formation.

     Observations of prion-like behavior from other systems may be informative for how scrapie is functioning. In yeast, two proteins have been described as prions, as they appear to exhibit prion-like behavior of nongenetic self-propagation. Ure2p is a yeast protein involved in the regulation of nitrogen metabolism. Ure2p appears to spontaneously convert to an inactive form, Ure2p* (Weissmann, 1994), exhibiting a phenotype referred to as [URE3]. The Ure2p* works to convert active Ure2p into inactive Ure2p*. [URE3] is a dominant trait and can not be explained by classical genetic principles. No nucleic acid has been found to associate with it, but it can be passed between cells by transfer of cytoplasm. Another non-mendelian prion-like factor in yeast is [psi+]. Like [URE3], [psi+] is sensitive to agents which affect the yeast stress response. In a search for mutlticopy suppressors of [psi+] function, a chaperonin was found which is able to modulate the presence of the [psi+] phenotype in cells (Chernoff, 1995). Hsp104 was found to be required for propagation of [psi+], yet overexpression of Hsp104 "cures" the cells of [psi+]. It is postulated that Hsp104 may act by fostering or enabling an intermediate folding state, which is then able to collapse into either the prion or non prion state. Hsp104 is part of the Clp family of genes conserved from prokaryotes to eukaryotes, including genes involved in the stress response of mammalian cells (Parsell et al., 1991).

     From all indications cited above, scrapie formation appears to be localized to subcellular vesicles. Examination of vesicles could address the following three questions: Where is the precise location of scrapie formation? Does the PrP dimer that has been postulated to be an important intermediate in scrapie formation exist in the location of scrapie formation? Does hsp73 co- localize or share any compartment with PrPC or PrPSc?


A succinct version in plain english

Back to Chris's page