The Effects of Trimethylamine N-Oxide on the Structural Stability of Prion Protein

Transmissible spongiform encephalopathies (TSEs) are diseases that affect the central nervous system in both humans and animals. TSEs can be ascribed to conformational conversion of prion protein (PrP). “Prion” stands for “proteinaceous infectious particle”, which was discovered in the disease-transmitting material of infectious brain tissues and named by Prusiner (Prusiner 1982). The normal, cellular form of PrP (PrPC) is a α-helix-rich glycoprotein attached to the outer cell surface by a glycophosphatidyl inositol linkage. The biological role of PrPC remains ambiguous. Since Cu2+ is unique among divalent metal ions in its ability to bind to PrPC in the octarepeat region (four copies of the repeat ProHisGlyGlyGlyTrpGlyGln) in the N-terminal domain (Whittal et al. 2000), the prion protein has been suggested to play a role in maintaining cellular copper concentration and signal transduction (Brown et al. 1998; Mouillet-Richard et al. 2000). PrPC is non-infectious, whereas the abnormal and infectious, Scrapie isoform of PrP (PrPSc) has been considered to be the major infectious component of the genetic, sporadic and transmissible fatal neuro degenerative prion diseases that affect both human and animals (Prusiner 1998). PrPSc is rich in β-sheet structures and it has a pronounced tendency to misfold and to subsequently aggregate into highly stable and insoluble amyloid plaques. This PrPSc plagess are resistant to digestion by proteinase K, whereas PrPC is sensitive to digestion by proteinase K (Prusiner 1998). PrPSc is responsible for many diseases in humans, including Kuru, Gerstmann-Sträussler-Scheinker disease (GSS), fatal familial insomnia (FFI), CreutzfeldtJakob disease (CJD) and the BSE-related, variant Creutzfeldt-Jakob disease (vCJD) (Horwich & Weissman 1997; Prusiner 1997; Will et al. 1996). The infectious nature of these fatal diseases differ from other infectious diseases in that the pathogen is a proteinaceous particle rather than typical pathogens, such as viruses, bacteria and fungi. Therefore, a “proteinonly“ hypothesis has been proposed (Griffith 1967; Prusiner 1998). A protein designated as “protein X“ has been proposed to be involved in structural conversion of PrPC into PrPSc (Telling et al. 1995) in the protein-only model. After decades of study on prion proteins, protein X has not been identified. Thus, the mechanism of how PrPC is converted into PrPSc remains ambiguous. The structure of prion proteins has been studied in human (Calzolai & Zahn 2003; Donne et al. 1997; Zahn et al. 2000), hamster (James et al. 1997) and mouse sequences (Gossert et al.

nervous system in both humans and animals.TSEs can be ascribed to conformational conversion of prion protein (PrP)."Prion" stands for "proteinaceous infectious particle", which was discovered in the disease-transmitting material of infectious brain tissues and named by Prusiner (Prusiner 1982).The normal, cellular form of PrP (PrP C ) is a α-helix-rich glycoprotein attached to the outer cell surface by a glycophosphatidyl inositol linkage.The biological role of PrP C remains ambiguous.Since Cu 2+ is unique among divalent metal ions in its ability to bind to PrP C in the octarepeat region (four copies of the repeat ProHisGlyGlyGlyTrpGlyGln) in the N-terminal domain (Whittal et al. 2000), the prion protein has been suggested to play a role in maintaining cellular copper concentration and signal transduction (Brown et al. 1998;Mouillet-Richard et al. 2000).PrP C is non-infectious, whereas the abnormal and infectious, Scrapie isoform of PrP (PrP Sc ) has been considered to be the major infectious component of the genetic, sporadic and transmissible fatal neuro degenerative prion diseases that affect both human and animals (Prusiner 1998).PrP Sc is rich in β-sheet structures and it has a pronounced tendency to misfold and to subsequently aggregate into highly stable and insoluble amyloid plaques.This PrP Sc plagess are resistant to digestion by proteinase K, whereas PrP C is sensitive to digestion by proteinase K (Prusiner 1998).PrP Sc is responsible for many diseases in humans, including Kuru, Gerstmann-Sträussler-Scheinker disease (GSS), fatal familial insomnia (FFI), Creutzfeldt-Jakob disease (CJD) and the BSE-related, variant Creutzfeldt-Jakob disease (vCJD) (Horwich & Weissman 1997;Prusiner 1997;Will et al. 1996).The infectious nature of these fatal diseases differ from other infectious diseases in that the pathogen is a proteinaceous particle rather than typical pathogens, such as viruses, bacteria and fungi.Therefore, a "proteinonly" hypothesis has been proposed (Griffith 1967;Prusiner 1998).A protein designated as "protein X" has been proposed to be involved in structural conversion of PrP C into PrP Sc (Telling et al. 1995) in the protein-only model.After decades of study on prion proteins, protein X has not been identified.Thus, the mechanism of how PrP C is converted into PrP Sc remains ambiguous.The structure of prion proteins has been studied in human (Calzolai & Zahn 2003;Donne et al. 1997;Zahn et al. 2000), hamster (James et al. 1997) and mouse sequences (Gossert et al. 2005;Riek et al. 1996).The structure of truncated mouse PrP (121-231) has been solved and reveals a α-helical protein containing 52% of α-helical and 3% of β-sheet structures (PDB code: 1AG2, (Riek et al. 1996)).Furthermore, NMR structure of hamster prion 90-231 reveals a flexible N-terminus and a structured C-terminal fragment (125-228) containing 43% of αhelical and 6% of β-sheet structures (PDB code: 1B10, (James et al. 1997)).The high flexibility of the N-terminus was confirmed in recombinant full-length hamster prion protein PrP 29-231 (Donne et al. 1997).The representative NMR structure of mouse prion is shown in Figure 1.The formation of amyloid fibrils can be described by the polymerization model (Tsong & Baldwin 1972) illustrated in Figure 2. In this nucleation-dependent polymerization model, the conversion of proteins into nuclei that subsequently serve as seeds for amyloid fibrils is the rate-limiting step.Therefore, amyloid formation has a lag phase in which nucleation from protein monomers takes place.The spectroscopic signal for detecting amyloid fibrils is very weak during this nucleation phase, but the signal will enhance significantly during the fast polymerization catalyzed by the nuclei in the growth phase.Finally, the signal remains constant at a steady stateas that the ordered amyloid fibrils and protein monomers are at equilibrium.In previous studies of recombinant PrP, a partially folded intermediate of PrP 90-231 has been suggested (Apetri et al. 2004;Kuwata et al. 2002;Nicholson et al. 2002).These partially unfolded states can potentially initiate amyloid formation (Abedini & Raleigh 2009;Chiti & Dobson 2009).Consequently, extensive works on the interaction of samll molecules and PrP have been conducted in order to find pharmacological chaperones that can stabilize the structure of PrP C (Nicoll et al. 2010).Chemical chaperones is an important target in the developments of therapies for prion disease.In the study of protein folding, osmolyte is a well-know factor that acts as a chemical chaperone to assist proteins to fold.Furthermore, cellular osmolarity can be changed by the ion-transportations.An increase in intracellular osmolarity caused by ion-transoprtation resulting astrocyte dysfunction in neurological disorders has been reviewed (Seifert et al. 2006).Therefore, it is essential to study osmolytes that affect folding of prion proteins.The chemical structure of the osmolyte trimethylamine N-oxide (TMAO) is illustrated in Figure 3 that TMAO is a zwitterionic compound at neutral pH, but is positively charged below pH 4.7 (Lin & Timasheff 1994).TMAO is a well known protective osmolyte that can increase structural stability of proteins against chemical or temperature-induced denaturation (Baskakov et al. 1998;Bolen & Baskakov 2001;Celinski & Scholtz 2002;Gursky 1999;Qu et al. 1998).TMAO has an extraordinary ability to force thermodynamically unstable proteins to fold (Baskakov & Bolen 1998).In contrast, TMAO destabilizes proteins at low pH (Singh et al. 2005).Fig. 3.Chemical structure of zwitterionic TMAO in equilibrium with its cationic form A previous study by Tatzel and co-workers indicated that TMAO and the organic solvent dimethylsulfoxide (DMSO) prevent scrapie formation in vitro (Tatzelt et al. 1996).This study suggests a potential stratagy for preventing PrP Sc by stabilizing the conformation of PrP C .In contrast, later studies indicated that TMAO destabilized the α-helical conformation of fulllength human prion at neutral pH at high temperature (Nandi et al. 2006) and human prion proteins at low pH (Granata et al. 2006).The structure of α-helix-rich prion protein is destabilized and the prion protein is converted to β-sheet structured oligomeric species (Nandi et al. 2006).

N-terminus C-terminus
In this study, we examined the effect of TMAO on conformational stability in full-length mouse prion proteins (MoPrPs) and their structural conversion to amyloid fibrils.A possible mechanism of TMAO on fibril formation is proposed based on the nucleation-dependent polymerization model.

Materials and methods
We characterized TMAO-induced structural change in MoPrP by circular dichroism spectroscopy and 8-anilino-1-naphthalenesulfonic acid (ANS)-binding fluorescence assay, and compared the effect of TMAO on the formation of amyloid fibrils by thioflavin T (ThT) binding assay and transmission electron microscopy (TEM).Chemicals used in this work were purchased from Sigma Chemcial Co. (St.Louis, Missouri, USA), and used without further purification.

Protein expression and purification
Plasmid pET101 encoding mouse prion (23-231) was transformed into competent E. coli BL21 (DE3) and expressed in the form of inclusion bodies.The protein was purified on a Nisepharose column according to a previously described procedure (Bocharova et al. 2005).The purity of the isolated protein was confirmed by SDS-PAGE.The correct native conformation, containing mainly α-helical structures, was confirmed by circular dichroism spectroscopy.

Far-UV circular dichroism spectroscopy
Far-UV circular dichroism spectra of mouse prion was recorded with a J-815 (Jasco, Japan) spectropolarimeter equipped with a Peltier temperatue control system (model PTC 423).For measurements in the far-UV region, a quartz cell with a path length of 0.1 cm was used in nitrogen atmosphere.For protein samples, the concentration of MoPrP was kept constant at 10 μM in 10 mM 2-(N-morpholino)ethanesulfonic acid (MES, pH 6.0).Far-UV circular dichroism spectra between 200 and 250 nm were recorded.An accumulation of five scans with a scan speed of 50 nm per minute was performed at 20 °C.The thermal-induced denaturation of proteins was conducted by heating protein solutions at the rate of 1 °C/min, and measuring the molar ellipticity at 222 nm every 0.5 °C.The reading of molar ellipticity at 222 nm in accordance with temperature change was normalized for further analysis.The data were analyzed with Origin 6.0 from OriginLab (Northampton, Massachusetts, USA).

ANS fluorescence assay
ANS solution was added into 10 μM of TMAO-incubated MoPrP in 10 mM MES (pH 6.0) to a final concentration of 100 μM.For each sample, two fluorescence emission spectra were collected by a F-4500 fluorometer (Hitachi, Japan) with excitation either at 295 nm or 385 nm, respectively.

Fibril formation
Fibril conversion was conducted at 20 µM protein in the solution containing 2 M guanidine hydrochloride (GdnHCl) and 50 mM MES (pH 6) at 37 °C as described in a previous study on mouse prion fibrils (Bocharova et al. 2005).At the end of the experiments, the fibril samples were dialyzed against 20 mM sodium acetate (pH 5.5) for further experiments.

ThT fluorescence assay of fibril conversion
Aliquots withdrawn during the time course of incubation at 37 °C were diluted into 5 mM sodium acetate buffer (pH 5.5) to a final sample concentration of 0.5 μM.ThT solution was added to a final concentration of 10 μM.For each sample, two emission spectra were collected by a F-4500 fluorometer (Hitachi, Japan) equipped with a 150 W Xenon lamp.The fluorescence spectra were recorded from 470 to 550 nm with excitation wavelength of 450 nm.The maximum fluorescence emission at 482 nm was determined.The reported fluorescence readings were average values for two measurements.

TEM
The aliquotes of mouse prion fibril samples were stained with 2% tungsten phosphoric acid onto carbon-coated 200-mesh copper grids.The samples were adsorbed onto copper grids for 30 sec and subsequently washed with PBS and mQH 2 O.The samples were air-dried before imaging.The TEM images were collected by a H-7100 TEM (Hitachi, Japan).The analysis of fibril length and width was carried out on WICF ImageJ software (National Institutes of Health).

Results and discussion
We characterized the effect of the osmolyte TMAO on the structure of MoPrP by far-UV circular dichroism spectroscopy and ANS-binding fluorescence assay.Furthermore, we tested the effect of TMAO on amyloid fibril formation.The prion fibrils grown at different concentrations of TMAO were analyzed by TEM.

Secondary structure and protein stability
We collected far-UV circular dichroism spectra for MoPrP at pH 6.0 at 20 °C.As shown in Figure 4, MoPrP has the typical feature of α-helical conformation as indicated by the circular dichroic signals at 208 nm and 222 nm.In prior investigations of TMAO and its effects on MoPrP structure, circular dichroism spectra of MoPrP in 2 M TMAO were recorded every hour.These spectra collected within 6 hours are very similar (data not shown).Since the extension of incubation does not affect the structural features of MoPrP, the following spectroscopic experiments were conducted without prolonged incubation.

TMAO causes unfolding and destabilizes MoPrP
We investigated the addition of TMAO in the range of 0.5 ~ 2 M and its effect on MoPrP using circular dichroism spectroscopy.The results are presented in Figure 4A.Osmolyte TMAO exhibits substantial interference below 205 nm.Therefore, the interfering circular dichroic signal was removed from the plot.In the absence of TMAO, MoPrP has clear circular dichroic feature of α-helical conformation at 208 and 222 nm.After addition of 0.5 M and 1.0 M TMAO, the circular dichroic signal indicative of α-helix structure at 208 nm decreased slightly, while the signal at 222 nm remained generally unchanged.Significantly, addition of 1.5 M and 2.0 M weakens the entire α-helical conformation.The unfolding of α-helical structure has been reported with similar work performed at low pH (Granata et al. 2006).Consistently, the significant unfolding of prion proteins is carried out at > 1.0 M TMAO as illustrated in Figure 4B.Differently, the dichroic signal at 208 nm is affected by TMAO more severely than that at 222 nm.The dichroic signal at 208 nm dominants the dichroic curve of α-helix in the absence of TMAO, but this signal is largely weakened by the addition of TMAO.The dichroic curve of TMAO-affected partially folded α-helical structure observed in this work is not identical to the curves observed in the similar work performed at high temperature (Nandi et al. 2006).These results indicated that TMAO affect the structure of prion proteins differently under different conditions.At neutral pH, TMAO is zwitterionic, but it becomes cationic at low pH.Full-length prion protein includes polybasic N-terminal region 23-30 which is essential for effective folding into the native cellular conformation (Ostapchenko et al. 2008).Therefore, the differentiation of TMAO-induced conformational change at different pH values can be partially, if not all, ascribed to the interaction between TMAO and the polybasic N-terminal region.TMAO is known to increase the melting temperature of various proteins (Arakawa & Timasheff 1985).Therefore, in addition to the conformation change, the structural stability of MoPrP in the presence of TMAO was examined with thermal-induced denaturation monitored by the molar ellipticity at 222 nm.The molar ellipticity was normalized to represent the fraction of unfolded proteins as presented in Figure 5.The curves of heat-induced denaturation can be described by a cooperative two-state model of denaturation in which proteins are transited between natively folded states and denatured unfolded states.By monitoring the circular dichroic signal at 222 nm, the natively folded MoPrP has the minimum intensity, whereas fully unfolded proteins has the maximum readings.These denaturation curves can be analyzed based on the Gibbs-Helmholtz equation: Based on the above equation, the melting temperature (T m ) of MoPrP can be determine.In the absence of TMAO, the T m of MoPrP is 68.63 °C.Addition of 0.5 M TMAO does not affect the thermal-denaturation curve significantly, as the T m value was calculated to be 68.19 °C.
In the presence of 1.0 ~ 1.5 M TMAO, the thermal-denaturation curve shifted toward low temperature significantly, as MoPrP starts to d e n a t u r e a t 3 2 °C .I n c o n t r a s t t o t h e denaturation curve collected in the absence of TMAO, the thermal-denaturation curves of MoPrP treated with ≥ 1.0 M TMAO are not typical of cooperative two-state unfolding.These curves implies that TMAO changes the unfolding pathway, and that more intermediates are present at high concentration of TMAO.

TMAO-induced conformational change within the native state
The hydrophobic dye ANS has been widely used to detect the conformational changes in proteins, as ANS preferentially binds to hydrophobic protein surfaces  The fluorescence spectra of ANS in the presence of TMAO illustrate a strong peak at 408 nm, a medium, broad peak at ~460 nm and a weak, broad peak at ~490 nm overlapped with the broad peak at 460 nm.The fluorescence at 490 nm is a typical emission of ANS-binding as observed in Figure 6.This weak signal at 490 nm could be ascribed to MoPrP-bound ANS molecules that do not accept energy from Trp efficiently.As the addition of TMAO causes the peaks at 408 nm and at 460 nm to increase at different degrees, it is predicted that there are two distinct hydrophobic pockets near the N-terminal region but away from Trp with different distances.The formation of amyloid fibrils in the presence of 0 ~ 1.5 M TMAO follows the nucleationdependent polymerization model as illustrated in Figure 2. According to this nucleationdependent polymerization model, the lag phase of amyloid formation from MoPrP can be determined by fitting the time-dependent changes in the ThT fluorescence (F) over time (t) of the reaction as shown in the following equation (Bocharova et al. 2005 Where F 0 is the minimum level of ThT fluorescence during the lag phase, ΔF is the difference of ThT fluorescence between the maximum level (steady state) and the minimum level (lag phase), k is the rate constant of fibril growth (h -1 ), and t m is the observed time at the midpoint of transition.The lag time (t l ) of fibril formation can be calculated as: According to these two equations, the lag time for fibril formation in the presence of 0, 0.5 and 1.0 M TMAO was determined as presented in Figure 9B.Previous work on fibril conversion from α-synuclein in the presence of TMAO reported acceleration of amyloid formation in a wide range of TMAO concentration (1.0 ~ 3.0 M) due to osmolyte-induced stabilization of the partially folded intermediate (Uversky et al. 2001).Similarly, TMAO stabilizes some partially folded intermediates.However, these partially folded intermediates slightly disfavour the nucleation reaction resulting in slight extension of the nucleation phase and also the following growth phase.

Morphology of prion fibrils grown with TMAO
Variation of conditions for fibril formation from the identical protein can yield different morphology.For example, Makarava and Baskakov have observed curved and straight fibrils from the same mouse prion protein and identical solvent conditions but different in the shaking speed (Makarava & Baskakov 2008).Therefore, we investigated the morphology of mouse prion fibrils by electron microscopy.In our experiment, for mature mouse prion fibrils grown in 2 M GdnHCl, the length is mostly longer than 500 nm, and the width is about 20 nm as illustrated in Figure 10.The growth phase of fibril conversion in the absence of TMAO is very short, whereas this elongation process takes longer in the addition of TMAO.To compare the morphology of prion fibrils during elongation in the absence and in the presence of TMAO, we collected TEM images of fibril samples in the middle point of growth phase.As illustrated in Figure 11, the prion fibrils grown in the absence of TMAO are long and straight, whereas fibrils converted in the presence of TMAO look like short rods.Increasing TMAO to 1.0 M shortens the prion fibrils, so that the fibril length ranges from 20 to 140 nm with the most populated length being 40 ~ 60 nm.The average length of prion fibrils is 144 nm in the absence of TMAO.Addition of TMAO shortens the average length of prion fibrils to 92 and 61 nm in the presence of 0.5 M and 1.0 M, respectively.In contrast to the fibril length, the width of prion fibrils is similar as illustrated in Figure 12B, and the average of the width is 19 nm regardless the presence of TMAO.
The structural effect of osmolyte TMAO on prion proteins and fibril conversion has been investigated extensively in this work.At pH 6.0, 20 °C, the α-helical conformation of MoPrP is unfolded by TMAO.When the concentration of TMAO is low at ≤ 1.0 M, the degree of unfolding is small and the fibril conversion takes place on the mostly folded α-helical structures.In contrast to low concentration of TMAO, high concentration of TMAO at >1.0 M largely unfolds the α-helix and this unfolding inhibits amyloid formation completely.In other words, low concentration of TMAO decellarates the prion nucleation represented by lag phase in nucleation-dependent polymerization model, whereas TMAO enhances the growth of prion fibrils.This contrast effect indicates that the structural requirement of prion proteins in nucleation and in fibril elongation is different.The importance of partial unfolding of prion in amyloid formation has been suggested (Apetri et al. 2004;Kuwata et al. 2002;Nicholson et al. 2002).The structure of partially folded prion protein that initiates amyloid formation remains unclear.Our recent computational simulation proposed that the ).In prion nucleation, high α-helical conformation is preferred and intermolecular β-sheet interactions occur between prion monomers (Lee et al. 2010).This interaction of intermolecular β-sheets is most likely the initiation of nucleation.This intermolecular interaction might be correlated with a double-layered filament within one fibril as observed by atomic force microscopy (Anderson et al. 2006).When the α-helical structure is mostly lost and the intermolecular interaction is largely weakened, the amyloid formation can not take place due to failure of nucleation as illustrated in Figure 13.After nuclei are formed as fibril seeds, the elongation requires stable interactions between nuclei and proteins.As amyloid fibrils have been proposed to have hydrophobic interior (Buchete & Hummer 2007;Carroll et al. 2006), the area of the hydrophobic surface is important in fibril elongation.The hydrophobicity at the ANS-binding pockets that accept energy from Trp is induced significantly by TMAO, and the high hydrophobicity promotes hydrophobic interactions of nuclei and proteins to form the hydrophobic interior.Notably, in the presence of low concentration of TMAO, the amount of fibrils is significantly increased, but the fibril length is shortened in the middle of growth phase.This phenomenon could be due to adjustment of interaction between nuclei under high hydrophobicity as judged by long growth phase in the presence of TMAO.Taking togther these findings, a certain amount of α-helical conformation seems to be required in the structural conversion of prion proteins in nucleation, and hydrophobic interaction is essential in fibril elongation.

Conclusion
In this study, we have investigated the structural effect of osmolytes TMAO on MoPrP proteins and the features of mouse prion fibrils.Low concentration of TMAO alters α-helical structures resulting deceleration of nucleation serving as fibril seeds, but TMAO enhances hydrophobicity promoting fibril growth.High concentration of TMAO inhibits amyloid formation completely by inducing loss of α-helical conformation.This study provides more information in the details of molecular structure and interaction in amyloid formation.These findings are essential in further search for therapeutic therapies for prion disease.

Fig. 1 .
Fig. 1.The NMR structure of mouse prion protein (PDB code: 1AG2, (Riek et al. 1996)).The α-helical structures are represented by red ribbons, and the antiparallel β-sheet structure is colored in yellow.This figure was produced using the PyMol graphic package

Fig. 4 .
Fig. 4. (A) Far-UV circular dichroism spectra of 10 μM MoPrP measured after the addition of TMAO.(B) Comparison of mollar ellipticity at 222 nm and 208 nm in various concentration of TMAO

Fig. 5 .
Fig. 5. Thermal-induced denaturation of MoPrP monitored the circular dichroic signal at 222 nm . The binding of ANS to hydrophobic area causes significant increase in the fluorescence emission of ANS.Therefore, we detected ANS fluorescence emission upon excitation at at 385 nm to detect the change of hydrophobicity of MoPrP as presented in Figure6.Clearly, ANS-bound MoPrP emits fluorescence at 470 nm and 495 nm upon excitation at 385 nm which indicates the presence of two distinct ANS-binding pockets.Addition of 0.5 M TMAO weakens the fluorescence emission of ANS slightly, but addition of 1.0 ~ 2.0 M TMAO substantially increased ANS fluorescence.The increase of ANS emission at 470 nm is more pronounced than the fluorescence at 495 nm.It indicates that the hydrophobicity represents by 470 nm is more sensitive to TMAO than the other one at 495 nm.Overall, this result indicates that TMAO increases hydrophobic surface at two distinct sites in MoPrP.

Fig. 6 .
Fig. 6.Fluorescence of ANS in the absence and in the presence of TMAO upon excitation at 385 nm

Fig. 7 .
Fig. 7. Fluorescence of ANS in the absence and in the presence of TMAO upon excitation at 295 nm

Fig. 8 .
Fig. 8. Kinetic traces for amyloid fibril formation from MoPrP monitored by ThT-binding fluorescence in the presence of 2 M GdnHCl and various concentrations of TMAO at pH 6.0 at 37 °C.The fluorescence readings were average values of two measurementsThe quantity of prion fibrils judged by the maximum ThT fluorescence in the presence of TMAO is presented in Figure9A.Clearly, addition of 0.5 and 1.0 M TMAO promotes amyloid formation as the reading of ThT-fluorescence is greatly enhanced in comparison to that in the absence of TMAO.In contrast, when the concentration of TMAO was increased further, the fibril conversion was inhibited.It is interesting to observe enhancement of fibril conversion by TMAO at ≤ 1.0 M.This result is consistent with previous work performed on α-synuclein that 1.0 M TMAO yields large amount of fibril whereas 2.0 and 3.0 M of TMAO inhibits the fibril formation(Uversky et al. 2001).

Fig. 9 .
Fig. 9. (A)The quantity of amyloid fibrils converted from MoPrP monitored by the maximum of ThT-binding fluorescence measured two times, and (B) the calculated lag phase of fibril conversion in the absence and in the presence of TMAO

Fig. 10 .
Fig. 10.TEM image of mature prion fibrils grown in 2 M GdnHCl at pH 6.0 at 37 °C.The scale bar represents 500 nm

Fig. 11 .
Fig. 11.TEM images of mouse prion fibrils at the middle point of growth phase (A) in the absence of TMAO and (B) in the presence of 1.0 M TMAO.The scale bar represents 200 nmOligomers, protofibrils and mature fibrils involved in the formation of amyloid fibrils are typically distinguished by the length and width of the fibrils.Therefore, we statistically compared the length and the width of prion fibrils converted in the absence and in the presence of 0.5 M and 1.0 M TMAO.As illustrated in Figure12A, in the absence of TMAO, the fibril length is populated in two groups including one normal distribution between 60 ~ 180 nm and an elongation at > 180 nm.Prion fibrils with the addition of 0.5 M TMAO are normally distributed between 20 ~ 160 nm and the most populated length is 80 ~ 120 nm.Increasing TMAO to 1.0 M shortens the prion fibrils, so that the fibril length ranges from 20 to 140 nm with the most populated length being 40 ~ 60 nm.The average length of prion fibrils is 144 nm in the absence of TMAO.Addition of TMAO shortens the average length of prion fibrils to 92 and 61 nm in the presence of 0.5 M and 1.0 M, respectively.In contrast to the fibril length, the width of prion fibrils is similar as illustrated in Figure12B, and the average of the width is 19 nm regardless the presence of TMAO.The structural effect of osmolyte TMAO on prion proteins and fibril conversion has been investigated extensively in this work.At pH 6.0, 20 °C, the α-helical conformation of MoPrP is unfolded by TMAO.When the concentration of TMAO is low at ≤ 1.0 M, the degree of unfolding is small and the fibril conversion takes place on the mostly folded α-helical structures.In contrast to low concentration of TMAO, high concentration of TMAO at >1.0 M largely unfolds the α-helix and this unfolding inhibits amyloid formation completely.In other words, low concentration of TMAO decellarates the prion nucleation represented by lag phase in nucleation-dependent polymerization model, whereas TMAO enhances the growth of prion fibrils.This contrast effect indicates that the structural requirement of prion proteins in nucleation and in fibril elongation is different.The importance of partial unfolding of prion in amyloid formation has been suggested(Apetri et al. 2004;Kuwata et al. 2002;Nicholson et al. 2002).The structure of partially folded prion protein that initiates amyloid formation remains unclear.Our recent computational simulation proposed that the

Fig. 12 .
Fig. 12. Histrogram of length and width of prion fibrils measured on 100 fibrils in the absence and in the presence of 0.5 M and 1.0 M TMAO