Disease-associated F198S Mutation Increases the Propensity of the Recombinant Prion Protein for Conformational Conversion to Scrapie-like Form*

The critical step in the pathogenesis of transmissible spongiform encephalopathies (prion diseases) is the conversion of a cellular prion protein (PrPc) into a protease-resistant, β-sheet rich form (PrPSc). Although the disease transmission normally requires direct interaction between exogenous PrPSc and endogenous PrPC, the pathogenic process in hereditary prion diseases appears to develop spontaneously (i.e. not requiring infection with exogenous PrPSc). To gain insight into the molecular basis of hereditary spongiform encephalopathies, we have characterized the biophysical properties of the recombinant human prion protein variant containing the mutation (Phe198 → Ser) associated with familial Gerstmann-Straussler-Scheinker disease. Compared with the wild-type protein, the F198S variant shows a dramatically increased propensity to self-associate into β-sheet-rich oligomers. In a guanidine HCl-containing buffer, the transition of the F198S variant from a normal α-helical conformation into an oligomeric β-sheet structure is about 50 times faster than that of the wild-type protein. Importantly, in contrast to the wild-type PrP, the mutant protein undergoes a spontaneous conversion to oligomeric β-sheet structure even in the absence of guanidine HCl or any other denaturants. In addition to β-sheet structure, the oligomeric form of the protein is characterized by partial resistance to proteinase K digestion, affinity for amyloid-specific dye, thioflavine T, and fibrillar morphology. The increased propensity of the F198S variant to undergo a conversion to a PrPSc-like form correlates with a markedly decreased thermodynamic stability of the native α-helical conformer of the mutant protein. This correlation supports the notion that partially unfolded intermediates may be involved in conformational conversion of the prion protein.

The critical step in the pathogenesis of transmissible spongiform encephalopathies (prion diseases) is the conversion of a cellular prion protein (PrP c ) into a protease-resistant, ␤-sheet rich form (PrP Sc ). Although the disease transmission normally requires direct interaction between exogenous PrP Sc and endogenous PrP C , the pathogenic process in hereditary prion diseases appears to develop spontaneously (i.e. not requiring infection with exogenous PrP Sc ). To gain insight into the molecular basis of hereditary spongiform encephalopathies, we have characterized the biophysical properties of the recombinant human prion protein variant containing the mutation (Phe 198 3 Ser) associated with familial Gerstmann-Straussler-Scheinker disease. Compared with the wild-type protein, the F198S variant shows a dramatically increased propensity to self-associate into ␤-sheet-rich oligomers. In a guanidine HClcontaining buffer, the transition of the F198S variant from a normal ␣-helical conformation into an oligomeric ␤-sheet structure is about 50 times faster than that of the wild-type protein. Importantly, in contrast to the wildtype PrP, the mutant protein undergoes a spontaneous conversion to oligomeric ␤-sheet structure even in the absence of guanidine HCl or any other denaturants. In addition to ␤-sheet structure, the oligomeric form of the protein is characterized by partial resistance to proteinase K digestion, affinity for amyloid-specific dye, thioflavine T, and fibrillar morphology. The increased propensity of the F198S variant to undergo a conversion to a PrP Sc -like form correlates with a markedly decreased thermodynamic stability of the native ␣-helical conformer of the mutant protein. This correlation supports the notion that partially unfolded intermediates may be involved in conformational conversion of the prion protein.
Transmissible spongiform encephalopathies, or prion diseases, comprise a group of fatal neurodegenerative disorders that affect both animals and humans. These disorders include bovine spongiform encephalopathy in cattle, scrapie in sheep, chronic waste disorder in deer and elk, and kuru, Creutzfeldt-Jacob disease, fatal familial insomnia, and Gerstmann-Straus-sler-Scheinker disease (GSS) 1 in humans. Transmissible spongiform encephalopathies may arise sporadically, may be inherited (familial forms), or may be acquired by transmission of an infectious agent (1)(2)(3)(4).
The prion diseases are associated with the accumulation of an abnormal form (PrP Sc ) of host-derived cellular prion protein (PrP C ). According to the "protein-only" hypothesis, PrP Sc represents the sole component of the infectious prion agent (1,5). Although the ultimate proof for this hypothesis is still missing (6), the central role of PrP in the pathogenesis of prion diseases is strongly supported by a wealth of biochemical and genetic data (1)(2)(3)(4). Furthermore, unexpected support for the proteinonly model comes from studying prion-like phenomena in yeast. A series of recent studies on protein-based inheritance in these organisms has provided mechanistic "proof of principle," unequivocally demonstrating that as predicted by the prion hypothesis, proteins can act as infectious agents by causing self-propagating conformational changes (7)(8)(9). Cellular human prion protein (PrP C ) is a glycoprotein that contains a single disulfide bond, is N-glycosylated, and attached to the plasma membrane by a C-terminally linked glycosyl phosphatidylinositol anchor. Although covalent structures of PrP C and PrP Sc have been found to be identical (10), the two isoforms have dramatically different physiochemical properties. PrP C is monomeric and readily degradable by protolytic enzymes, whereas PrP Sc exists as an insoluble aggregate that shows dramatically increased resistance to digestion by proteinase K (1)(2)(3)(4)11). Furthermore, spectroscopic data indicate that the two prion protein isoforms are characterized by different secondary structure; PrP C is largely ␣-helical, whereas PrP Sc is rich in ␤-sheet structure (12)(13)(14).
One of the key arguments in favor of the protein-only model is the link between familial prion diseases and specific mutations in the gene coding for human prion protein (1)(2)(3)15). Over 20 mutations in the human PrP gene have been identified to date that segregate with familial Creuzfeldt-Jakob disease, GSS, or fatal familial insomnia (1)(2)(3)15). Because the PrP C 3 PrP Sc conversion in hereditary prion diseases appears to occur spontaneously (i.e. without the presence of an exogenous PrP Sc ), studies with prion protein variants containing familial mutations could provide important clues regarding the molecular mechanism of the pathogenic process in spongiform encephalopathies. In this study, we have characterized the biophysical properties of the F198S variant of the recombinant human prion protein. The F198S mutation, associated with familial GSS (15,16), results in a marked destabilization of the native ␣-helical conformation of PrP C . Remarkably, the present data show that, compared with the wild-type PrP, the mutant protein is characterized by a greatly increased propensity to undergo a conversion to an oligomeric ␤-sheet-rich form with physicochemical properties similar to those of brain PrP Sc .

EXPERIMENTAL PROCEDURES
Plasmid Construction and Protein Purification-The plasmid encoding huPrP90 -231 with a N-terminal linker containing a His 6 tail and a thrombin cleavage site was described previously (17). The F198S mutant was constructed by site-directed mutagenesis using appropriate primers and a QuikChange kit (Stratagene). The wild-type and F198S variant proteins were expressed, cleaved with thrombin, and purified as described previously (17,18), with the exception that cleavage of F198S huPrP90 -231 was carried on for 48 h at room temperature in 10 mM acetate buffer, pH 5.6 (10 units of thrombin/mg of protein). The protein concentration was determined spectrophotometrically using the molar extinction coefficient (⑀ 276 ) of 21,640 M Ϫ1 cm Ϫ1 .
CD Measurements-The far UV CD spectra were obtained at room temperature on a Jasco J-810 spectropolarimeter. Typically, five spectra were averaged to improve the signal-to-noise ratio, and the spectra were corrected for small contribution of a buffer. The measurements were performed at room temperature in 0.2-mm-or 1-cm-pathlength cells. The urea-induced unfolding curves of the proteins were obtained using a model JWATS-489 automatic titrator attached to the spectropolarimeter. In a typical equilibrium unfolding experiment, two stock solutions of huPrP90 -231 at identical protein concentrations (1.2 M) were prepared: one in buffer alone (native protein) and one in an appropriate buffer containing 9 M urea (unfolded protein). The buffers used were 50 mM sodium acetate (pH 4, 5, and 5.5) and 50 mM potassium phosphate (pH 7). For each unfolding/refolding curve, a sample in 9 M urea was titrated (in ϳ0.1 M urea increments) to a sample of native protein in a 1-cm-pathlength cell. Upon each urea addition, the mixture was incubated for 10 s, and the ellipticity at 222 nm was recorded for 32 s. In the control experiments, it was verified that an incubation time of less than 1 s is sufficient for the system to reach equilibrium (the unfolding/refolding of the prion protein is very fast, occurring on the millisecond time scale) (19) and that the unfolding and refolding reactions are fully reversible. Urea was deionized using a mixture of anion exchange (trimethylbenzylammonium) and cation exchange (Dowex MR-3) resins, and its concentration was determined by measuring the refractive index.
Fourier Transform Infrared Spectroscopy-Infrared spectra were obtained at 25°C on a Bruker IFS 66 instrument. Samples of either wild-type huPrP90 -231 or F198S huPrP90 -231 were dissolved in a buffer prepared in D 2 O (50 mM acetate, pD 5.5) and placed between two calcium fluoride windows separated by a 50-m Teflon spacer. For each spectrum, 256 scans were accumulated, and the spectra were corrected for absorption of the buffer.
Size Exclusion Chromatography-Analytical size exclusion experiments were performed on a Bio-Sil SEC-250 gel filtration column (Bio-Rad) attached to a Shimadzu SCL-10Avp high pressure liquid chromatography system. The flow rate was 1.0 ml/min, and the elution of proteins was monitored by absorbance at 280 nm. The column was calibrated with molecular mass standards for size exclusion chromatography.
Dynamic Light Scattering-The light scattering experiments were performed at room temperature on a DynoPro-801 dynamic light scattering instrument (Protein Solution, Inc.). Prior to the measurements, the buffers were filtered through a 0.2-m membrane. The data were analyzed with Dynamics 4.0 software (Protein Solutions, Inc.).
Thioflavine T Assay-The progress of amyloid fibril formation was followed by a fluorimetric thioflavine T (ThT) assay (20). Typically, the proteins (60 or 120 M) were incubated at 37°C in 50 mM acetate buffer, pH 5.5. Small aliquots of each sample (5 or 10 l for 120 or 60 M protein solutions, respectively) were withdrawn at different time points and transferred to a quartz cell containing 10 M ThT in 50 mM potassium phosphate, pH 6.0. After 30 s of incubation at room temperature, the fluorescence was measured at 482 nm using the excitation wavelength of 450 nm. The final protein concentration in the ThT buffer was 3 M.
Electron Microscopy-Samples for electron microscopy were prepared by incubating the wild-type huPrP90 -231 or F198S huPrP90 -231 (120 M) at 37°C in a 50 mM sodium acetate buffer, pH 5.5. Following incubation for 3-5 days, a drop of each sample was placed on a carbon-coated 600-mesh copper grid (Electron Microscope Sciences) and negatively stained with 2% aqueous (w/v) uranyl acetate. Grid preparations were visualized using a JEOL 1200 transmission microscope operating at 80 keV.
Proteinase K Sensitivity-Samples of the wild-type huPrP90 -231 and the F198S variant (120 M in each case) were preincubated for 5 days at 37°C in 50 mM acetate, pH 5.5. The protein samples were then transferred to the digestion buffer (50 mM acetate, pH 5.5) containing various amounts of proteinase K (0 -50 g/ml) and incubated for 1 h at 37°C. The final protein concentration in the digestion mixture was 30 M. The reaction was terminated by boiling the samples in Laemmli electrophoretic buffer (Sigma). The protein fragments were then separated by SDS-polyacrylamide gel electrophoresis and detected by Coomassie Blue staining.

pH-dependent Thermodynamic Stability of the Wild-type huPrP90 -231 and the F198S Variant-
The effect of the F198S substitution on the thermodynamic stability of the prion protein at different pH values was probed by equilibrium unfolding in urea. At a given pH, the unfolding of the mutant protein occurs at a lower urea concentration than that of the wild-type huPrP90 -231 (unfolding curves not shown for brevity). In each case, the experimental curves could be fitted according to a two-state unfolding model (21). The results of such fitting analysis, expressed in terms of the midpoint unfolding urea concentration (C m ) and an apparent free energy difference between the native and unfolded state (⌬G o ) are summarized in Table I. It should be noted that the apparently good mathematical fit may be fortuitous because the equilibrium unfolding data alone are insufficient to prove the applicability of the two-state unfolding model. Nevertheless, the urea unfolding curves clearly indicate that at each pH value studied the Phe 198 3 Ser substitution results in a markedly decreased global thermodynamic stability of the prion protein.
Effect of F198S Mutation on GdnHCl-induced Conformational Conversion of huPrP90 -231-It has been recently shown that under mildly acidic conditions in the presence of low concentrations of GdnHCl (22,23) or urea and NaCl (24,25), the recombinant prion protein undergoes a transition to an oligomeric ␤-sheet-rich structure with essential physiochemical properties similar to those of brain PrP Sc . Here, we have examined the effect of the disease-associated F198S mutation on the conversion of huPrP90 -231 from the monomeric ␣-helical conformation to the oligomeric PrP Sc -like form. Fig. 1 shows the far UV CD spectra for the wild-type huPrP90 -231 (panel A) and the F198S variant (panel B) in 50 mM acetate buffer, pH 5.0, at a protein concentration of 95 M. In the absence of the denaturant, the spectra of both proteins have a double minimum at 222 and 208 nm, characteristic of ␣-helical structure. Consistent with the previous report (22), upon the addition of 1 M GdnHCl to the wild-type protein, there is a time-dependent decrease in negative ellipticity, concomitant with a change in the shape of the spectrum to one characteristic of ␤-sheet structure (minimum around 215 nm). Un- der the present experimental conditions, the ␣-helix 3 ␤-sheet transition for the wild-type huPrP90 -231 was relatively slow, with an apparent half-time of ϳ19 h. Remarkably, identical treatment of the F198S variant resulted in a very rapid transition to a ␤-sheet-rich structure. In the later case, the reaction was approximately 50 times faster compared with the wildtype protein; it was completed within less than 2 h and characterized by an apparent half-time of ϳ22 min ( Fig. 2A). The oligomerization state of the wild-type huPrP90 -231 and the F198S variant was studied by quasi-elastic light scattering and size exclusion chromatography. In the absence of GdnHCl, quasi-elastic light scattering data for freshly prepared solutions of both proteins (pH 5 buffer, room temperature) indicated the presence of single species corresponding to monomers with an apparent Stokes radius of ϳ2 nm. Upon the addition of 1 M GdnHCl, light scattering measurements indicated the presence of large protein oligomers (data not shown). The time course of the oligomerization reaction for both proteins could be followed by size exclusion chromatography. Representative elution profiles for the wild-type huPrP90 -231 and the F198S mutant following 150 min of preincubation in the presence of 1 M GdnHCl are shown in Fig. 2C. The chromatograms for both proteins show a peak corresponding to the monomer (elution volume of ϳ9.6 ml) and peak(s) corresponding to oligomeric species with an apparent molecular mass ranging from ϳ400 kDa to that higher than the fractionation limit of the column (700 kDa for globular proteins). For a given preincubation time, the proportion of the oligomeric protein was markedly higher for the F198S variant as compared with that for the wild-type protein. The time course of the loss of the peak corresponding to monomeric species of both proteins (Fig. 2B) indicates that, under the present experimental conditions, the oligomerization of the F198S variant is ϳ50 times faster than that of the wild-type protein (apparent half-times of 22 min and 19 h, respectively). For both the wild-type and mutant proteins, the kinetics of oligomer formation was very similar to the time course of changes in CD spectra (Fig. 2), indicating that the ␣-helix 3 ␤-sheet transition is intimately associated with selfassociation of the prion protein variants.
Spontaneous Conversion of F198S huPrP90 -231 to Oligomeric ␤-Sheet Form-A dramatically increased propensity of the F198S huPrP90 -231 variant to undergo a transition to an oligomeric ␤-sheet structure in the presence of GdnHCl prompted us to examine the behavior of the mutant protein in the absence of any chemical denaturants. Fig. 3 shows the time evolution of far UV CD spectra for the wild-type huPrP90 -231  and the F198S variant upon incubation at 37°C in 50 mM acetate buffer, pH 5.5. Clearly, the wild-type protein did not undergo any time-dependent conformational changes, remaining for many days in its native ␣-helical form (Fig. 3A). By contrast, under identical conditions the ␣-helical form of the F198S variant was highly unstable, undergoing a spontaneous transition to a ␤-sheet-rich structure (Fig. 3B). The kinetics of this reaction was dependent on the concentration of the protein; for example, at a F198S huPrP90 -231 concentration of 60 M, the apparent half-time of the transition (as measured by changes in CD spectra) was ϳ35 h, whereas at the concentration of 120 M, the half-time was reduced to 14 h.
Time-dependent changes in the secondary structure of F198S huPrP90 -231 were further examined by infrared spectroscopy. As shown in Fig. 4, the spectrum of a freshly prepared sample of the F198S variant has a maximum at 1648 cm Ϫ1 characteristic of a native ␣-helix-rich conformation. Upon incubation for 48 h at 37°C in deuterated sodium acetate buffer, pD 5.5, the spectrum of the protein shows two additional bands: one at 1618 cm Ϫ1 and one at 1685 cm Ϫ1 . The latter two bands are highly characteristic of an oligomeric ␤-sheet structure (26). Consistent with the CD data, under the same experimental conditions the infrared spectrum of the wild-type huPrP90 -231 did not undergo any time-dependent changes. The latter protein retains the native ␣-helical structure even after prolonged (many days) incubation in the GdnHCl-free buffer.
Similar to the conformational conversion in the presence of GdnHCl, the ␣-helix 3 ␤-sheet transition of F198S huPrP90 -231 in the denaturant free buffer was intimately associated with protein oligomerization. The nature of these oligomers was probed by the thioflavine T assay. This assay is based on the increase in ThT fluorescence upon binding of the dye to amyloid fibrils (20). As shown in Fig. 5, incubation of F198S huPrP90 -231 in 50 mM acetate buffer (pH 5.5) at 37°C results in a time-dependent increase in ThT binding, suggesting that the protein self-assembles into amyloid (or amyloid-like) structures. The kinetics of the formation of ThT-positive structures is concentration-dependent, becoming faster as the concentration of the protein is increased. Under the given experimental conditions, the time course of the ThT fluorescence increase was very similar to that of ␤-sheet structure formation (Fig. 5). Importantly, under identical conditions, the wild-type huPrP90 -231 did not exhibit any binding of the ThT dye. No ThT fluorescence increase was observed in the presence of the latter protein even after 7 days of incubation at 37°C.
The morphology of oligomeric structures formed by F198S huPrP90 -231 was examined by transmission electron microscopy. The micrographs of the protein preincubated at 37°C in acetate buffer, pH 5.5, showed ordered fibril-like structures with a diameter of 10 -20 nm and of variable length (Fig. 6). In addition to these ordered structures, irregular aggregates of variable shape and size were also observed. Consistent with the ThT data, no fibrils (or other oligomeric structures) could be detected under similar experimental conditions for the wildtype huPrP90 -231.
One of the hallmarks of brain PrP Sc is an unusually high resistance to digestion with proteinase K (1, 2, 11). The pro-   The data for the wild-type huPrP90 -231 (120 M) are represented by q. The kinetics of ␣-helix-␤-sheet transition for F198S huPrP90 -231 (120 M), as followed by changes in ellipticity at 222 nm, is represented by E. In each case, the proteins were incubated (without stirring or shaking) in 50 mM acetate buffer, pH 5.5, at 37°C. tease-resistant core of PrP Sc corresponds to the C-terminal region starting near residue 90. As shown in Fig. 7, treatment of a monomeric F198S huPrP90 -231 (or the wild-type huPrP90 -231) with as little as 1 g/ml of proteinase K resulted in a degradation of the protein. This is in contrast to the behavior of F198S huPrP90 -231 that was converted, by incubation at 37°C, to an oligomeric ␤-sheet structure. The latter protein displayed a markedly increased resistance to proteinase K digestion; the band corresponding to a ϳ17-kDa protein could be observed even upon treatment with as much as 10 g/ml of the enzyme. Similar incubation of the wild-type huPrP90 -231 did not result in the formation of a proteinase K-resistant species.
Most of the GdnHCl-independent PrP conversion experiments described in this study were performed in 50 mM sodium acetate buffer, pH 5.5. However, it should be noted that the spontaneous (i.e. not requiring any denaturant) transition of the F198S variant to an oligomeric form characterized by ␤-sheet-structure, proteinase K resistance, and affinity for the ThT dye could also be accomplished in a variety of other buffers, including 50 mM potassium phosphate, pH 6.0, and 0.1 M Tris-HCl, pH 6.8. Importantly, under identical buffer conditions and within the time scale of the experiments, the wildtype prion protein remained monomeric, readily degradable by proteinase K, and rich in ␣-helical structure. DISCUSSION One of the most intriguing aspects of prion diseases is the link between mutations in the PRNP gene and hereditary forms of Creuzfeldt-Jakob disease, GSS, and fatal familial insomnia. The molecular basis of this link is, however, unknown. Within the context of the protein-only model, it has been hypothesized that the familial mutations could facilitate the PrP C 3 PrP Sc conversion by the mechanism involving thermodynamic destabilization of the native ␣-helical structure of PrP C (27). However, the experimental support for this hypothesis is still missing. Furthermore, recent data questioned the general validity of the thermodynamic model, indicating that there is a great variability in the effect of individual disease-associated mutations on the stability of the cellular prion protein (28,29). Although some of these mutations indeed result in a significant destabilization of PrP C , the effects of others appear to be much more subtle. Clearly, more detailed studies are required to elucidate the relationship between mutation-induced perturbations in the biophysical and/or structural properties of prion protein and the molecular basis of hereditary prion disorders.
In the present study, we have focused on the human prion protein containing a Phe 198 3 Ser mutation. This substitution has been linked to GSS with extensive PrP amyloid deposits in the cerebrum, cerebellum, and midbrain (15,16). The key find-ing of the present study is that the recombinant prion protein variant F198S undergoes a time-dependent conversion to oligomeric structures that are ThT-positive and morphologically appear similar to those found in brain tissue of GSS patients (16). As in the case of brain deposits, the oligomers formed in vitro include fibrillar structures as well as less ordered aggregates. Importantly, the conversion in vitro of the mutant protein occurs spontaneously and under physiologically relevant buffer conditions. This is in contrast to the wild-type huPrP; conformational conversion of the latter protein could be observed only in the presence of denaturing agents (22)(23)(24)(25).
In addition to morphological similarities between F198S huPrP90 -231 aggregates and amyloid deposits found in GSS brain, the oligomeric form of the recombinant mutant protein shares many other physicochemical characteristics of a diseased brain PrP Sc , including high content of ␤-sheet structure and partial resistance to proteinase K digestion. These striking similarities notwithstanding, the question remains whether the recombinant F198S huPrP90 -231 oligomers described in this study are indeed structurally equivalent to the PrP Sc deposits characteristic of GSS with the F198S mutation. The amide I bands representing ␤-sheet structure in the infrared spectrum of F198S huPrP90 -231 oligomers (Fig. 4) appear at somewhat different positions than those reported for PrP Sc from scrapie-infected Syrian hamster brain tissue (12,13). This could potentially reflect subtle differences in the arrangement (or hydrogen bonding) of ␤-strands in the two proteins. However, these spectral differences could also be related to speciesand/or mutation-specific effects. Infrared spectroscopic data have been reported only for Syrian hamster PrP Sc ; no such data are yet available for human PrP Sc , let alone the F198S variant from the GSS brain.
Biological properties of the recombinant scrapie-like F198S huPrP90 -231 oligomers, including their potential neurotoxicity and infectivity, have not yet been tested. Questions regarding the infectivity of different types of PrP Sc and PrP Sc -like isoforms are of special interest and importance. In this context, one should note that attempts to infect experimental animals using brain tissue homogenates from F198S GSS patients have not been successful (15). It is thus possible that familial GSS with F198S mutation may represent a class of PrP-related amyloidosis that is different from transmissible forms of prion diseases. Clearly, further studies are needed to test the relationship between the oligomeric form of the recombinant PrP variant and authentic protein deposits found in F198S GSS brain. Nevertheless, the finding that a disease-related Phe 198 3 Ser mutation renders the recombinant prion protein highly susceptible to a conversion into a scrapie-like form is highly remarkable, bringing us a step closer toward understanding the molecular basis of familial prion disorders.
In particular, the present data provide new clues regarding the mechanism of prion protein conversion from ␣-helical conformation to ␤-sheet-rich oligomers. In contrast to the wildtype human PrP C (30,31), no high resolution structural data are yet available for the disease-associated F198S variant. Nevertheless, inspection of the wild-type structure indicates that the replacement of Phe 198 with a much smaller and polar Ser residue should result in a significant destabilization of the native protein structure (32). Urea unfolding experiments have indeed confirmed this effect, revealing drastic reduction in the thermodynamic stability for the F198S variant as compared with the wild-type PrP (see Ref. 29 and Table I). The link demonstrated in this study between mutation-dependent conformational conversion and thermodynamic stability of the prion protein is in line with observations for classical amyloidforming proteins such as transthyretin, immunoglobulin light chain, and lysozyme (33)(34)(35)(36). The pathway of amyloid formation by these proteins likely involves partially unfolded intermediates (33,(35)(36)(37). In the presence of destabilizing mutations, such intermediates would become more populated. Folding intermediates have also been postulated to play a key role in the PrP C 3 PrP Sc conversion (27). However, unlike for some other amyloidogenic proteins (33,(35)(36)(37), putative folding intermediates for PrP proved difficult to detect and characterize (19,38,39). This has led to speculation that it is the unfolded state of the prion protein that is directly converted to PrP Sc (19,38). Recent kinetic stopped flow data have, however, disputed this model, demonstrating that human prion protein folds by a three-state mechanism involving a monomeric folding intermediate (40). The role of this intermediate in the PrP C 3 PrP Sc conversion is indirectly indicated by experiments showing that the ␣-helix to oligomeric ␤-sheet transition of the recombinant human prion protein is strongly promoted in the presence of low and medium concentrations of urea, i.e. under conditions that could increase the population of partially unfolded intermediates. In contrast, conditions favoring the native state (no denaturant) or those that shift equilibrium toward the fully unfolded state (high concentrations of the denaturant) were not conducive to the conversion reaction (24). The present data for the F198S huPrP90 -231 variant further support and reinforce the notion that partial destabilization of the native state of PrP is critical to the transition to an oligomeric ␤-sheet-rich conformer. Although for the wild-type protein significant population of a conversion-conducive intermediate appears to require partially denaturing buffer conditions, in the case of the F198S variant a similar state may be populated under native conditions because of the destabilizing effect of the Phe 198 3 Ser mutation. Similar arguments may be applicable to other familial PrP variants carrying mutations that destabilize the native ␣-helical structure. It should be noted here that not all of the mutations linked to familial prion disorders have an intrinsic destabilizing effect on native PrP C structure relative to the fully unfolded state (28,29). However, even in the latter cases, mutations could potentially affect the stability and/or conformation of the intermediate state. Formation of conversion-conducive intermediate(s) in vivo could be further promoted by localization of the protein in specific compartments and/or mutation-dependent abnormalities in interactions with membranes, chaperones, or other components of the cellular environment. The notion that the partially unfolded state is critical to the spontaneous conversion of the recombinant PrP into a scrapie-like form is in line with the observations that partially destabilizing conditions (elevated temperature and the presence of chaotropes or chaperones) stimulate the PrP Sc -induced conversion of PrP C to a proteinase K-resistant form in a cell-free conversion assay (41)(42)(43).