Structural basis for the complete resistance of the human prion protein mutant G127V to prion disease

Prion diseases are caused by the propagation of misfolded cellular prion proteins (PrPs). A completely prion disease-resistant genotype, V127M129, has been identified in Papua New Guinea and verified in transgenic mice. To disclose the structural basis of the disease-resistant effect of the G127V mutant, we determined and compared the structural and dynamic features of the G127V-mutated human PrP (residues 91–231) and the wild-type PrP in solution. HuPrP(G127V) contains α1, α2 and α3 helices and a stretch-strand (SS) pattern comprising residues Tyr128-Gly131 (SS1) and Val161-Arg164 (SS2), with extending atomic distances between the SS1 and SS2 strands, and a structural rearrangement of the Tyr128 side chain due to steric hindrance of the larger hydrophobic side chain of Val127. The extended α1 helix gets closer to the α2 and α3 helices. NMR dynamics analysis revealed that Tyr128, Gly131 and Tyr163 underwent significant conformational exchanges. Molecular dynamics simulations suggest that HuPrP(G127V) prevents the formation of stable β-sheets and dimers. Unique structural and dynamic features potentially inhibit the conformational conversion of the G127V mutant. This work is beneficial for understanding the molecular mechanisms underlying the complete resistance of the G127V mutant to prion disease and for developing new therapeutics for prion disease.


Solution Structures of HuPrP(G127V) and WT HuPrP at pH 4.5.
Based on the resonance assignments and experimentally conformational restraints, we determined the solution structures of HuPrP(G127V) (PDB ID: 5YJ4) and WT HuPrP (PDB ID: 5YJ5) at pH 4.5 and 298K. The complete structural statistics are summarized in Table S1.
In HuPrP(G127V), the hydrophilic side chain of Tyr128 is rotated sharply so that its dihedral angle, chi (N-Cα-Cβ-Cγ), is reduced from 90° to 0° in WT HuPrP. This structural rearrangement might be introduced by the steric hindrance of the relatively larger hydrophobic side chain of Val127. This rotation pushes the phenyl ring of Tyr128 away from Ile182 and brings it closer to Gln186 (Fig. 2c,d). Although the HN Met129 -HN Tyr163 distance is identical for both structures (2.6 ± 0.1 Å), the Hα Leu130 -Hα Tyr162 distance is larger in the mutant (3.7 ± 0.2 Å vs. 2.2 ± 0.2 Å), as shown in Fig. 2e,f. The backbone atomic distances between SS1 and SS2 or between β1 and β2 are summarized in Table S2.
In the mutant protein, the C-terminal end of the α1 helix is extended to Arg156 (Figs 2a,b and S2a,b). In this configuration, the Arg156 side chain is closer to Thr190 and Thr191 at the C-terminal end of the α2 helix (Fig. S2c,d). In the α1-SS2 loop of HuPrP(G127V), the dihedral angle psi(N-Cα-C-N) of Tyr157 is nearly 60°, causing the pyrrolidine of Pro158 to retroflex approximately 180° (Fig. 2g,h). This alteration changes the atomic distances between Tyr157, Pro158, Val209 and Val210. Tyr157 becomes close to two α3-located residues, Val209 and Val210, and Pro158 moves away from Val209 and Val210. The retroflexion of the Pro158 side chain increases the curvature of the α1-SS2 loop in the mutant protein compared with the α1-β2 loop (His155-Gln160) in the WT (Fig. 2g,h). Additionally, atomic distances between SS2 and the disulfide bridge in the mutant are shorter than those between β2 and the disulfide bridge in the WT (Fig. S2g,h).
Furthermore, the G127V mutant also leads to a redistribution of the surface electrostatic potentials of the protein (Fig. S3). HuPrP(G127V) exhibits neutral potentials near residues Gly126-Ser135, while WT HuPrP shows positive potentials in this segment, except for Met129-Gly131. Additionally, compared with the WT protein, the mutant displays more positive potentials in the region near Arg146 and Arg151 of the α1 helix, and more negative potentials on the N-termini of the α1 and α3 helices.
The two different structures were calculated from their own different NOESY restraints originated from 3D 15 N-edited NOESY-HSQC and 13 C-edited NOESY-HSQC spectra (Fig. S4, Table S3). Several 1 H-1 H NOE peaks were missing, and many peaks were weaker in the SS segments from the HuPrP(G127V) than in those from the WT HuPrP (Fig. S4, Table S3). Furthermore, the structural differences were validated by backbone amide residual dipolar couplings (RDCs) measured from 2D 1 H-15 N IPAP-HSQC spectra (Figs S5 and S6). As indicated by the Q-values, the experimental RDCs from HuPrP(G127V) fitted better with the HuPrP(G127V) structure (Q = 0.532) than those with the WT HuPrP structure (Q = 0.798) and vice versa for the RDCs of the WT HuPrP (G127V vs WT: 0.856 vs 0.564) (Figs S5 and S6). In addition, the differences were confirmed with H/D exchanges based on 2D 1 H-15 N Fast-HSQC experiments (Figs S7 and S8). Remarkably, the amide proton of Gly131 in the SS segments from the HuPrP(G127V) was exchanged completely with D 2 O and became invisible in the HSQC spectrum than in those from the WT HuPrP (Figs S7 and S8). However, the amide protons of Met154 and His155 in the extended α1 helix from the HuPrP(G127V) became more stable than in those from the WT HuPrP (Figs S7 and S8).
Backbone amide relaxation analysis. To compare the dynamic features of the HuPrP(G127V) and WT HuPrP backbones, we performed a series of NMR relaxation experiments to obtain 15 N longitudinal relaxation rates (R 1 ), 15 N transverse relaxation rates (R 2 ) and { 1 H}-15 N heteronuclear steady-state NOEs ({ 1 H}-15 N NOEs) at two magnetic field strengths, 14.10 T and 19.97 T (Fig. 3). A total of 106 and 112 backbone amide resonances were used to analyse the dynamic features of the mutant and WT proteins. For HuPrP(G127V), residues in the N-terminal flexible segment did not show distinct differences in the R 1 and R 2 rates between the two magnetic fields but exhibited more negative { 1 H}-15 N NOEs at 14.10 T than those at 19.97 T. In contrast, the residues in the C-terminal structural core, except for those in the α2-α3 loop (Gly195-Thr199) and the C-terminus (Gly229-Ser231), displayed significant differences in the R 1 and R 2 rates and similar { 1 H}-15 N NOEs values (>0.6) between the two magnetic fields. The average R 1 rate at 14.10 T was larger than that at 19.97 T (1.1 s −1 vs. 0.8 s −1 ). All residues, except for Gly131 and Gln172 in the C-terminal structural core, displayed R 2 rates varying between 13.0 s −1 and 30 s −1 for both magnetic fields, with slightly higher values at 19.97 T. Furthermore, residues in the α2-α3 loop and the C-terminus showed larger R 1 rates and smaller R 2 rates as well as smaller { 1 H}-15 N NOEs. Overall, WT HuPrP showed R 1 rates, R 2 rates and { 1 H}-15 N NOEs roughly similar to HuPrP(G127V) for the two magnetic fields (Fig. 3). Both proteins exhibited larger differences in the R 2 / R 1 ratios between 14.10 T and 19.97 T.
Interestingly, the G127V mutant showed distinctly changed R 2 rates for the residues located in the SS1 and SS2 segments. The R 2 rates of the SS1-located residue Gly131 in the mutant protein were 38.5 s −1 at 14.10 T and 44.7 s −1 at 19.97 T, which were much larger than those in the WT (24.6 s −1 at 14.10 T and 33.4 s −1 at 19.97 T). Moreover, Gly131 also displayed significantly different R 2 /R 1 ratios between the two proteins. Furthermore, the R 2 rates of the SS2-located residue Tyr163 in the mutant protein were slightly larger than those in the WT (14.10 T: 21.8 s −1 . However, the R 2 rates for the SS1-located residues Met129 and Leu130 in the mutant protein were extremely similar to those in the WT. Regrettably, the relaxation data of the SS2-located residues Val161 and Tyr162 were not suitable for relaxation analysis because of resonance overlapping in the mutant protein. As expected, the G127V mutation more or less altered the R 2 rates of the α2 helix residues. Although the R 2 rate of Gln172 at 14.10 T was almost identical for both the mutant and WT proteins (28.5 s −1 vs. 27.6 s −1 ), this value at 19.97 T was smaller in the mutant than that in the WT (34.4 s −1 vs. 37.4 s −1 ). Furthermore, Ile182 in the mutant displayed slightly increased R 2 rates compared to that in the WT protein (24.9 s −1 vs. 22.9 s −1 at 14.10 T; 29.4 s −1 vs. 27.5 s −1 at 19.97 T). In addition, the R 2 rate of Gln186 in the mutant was much larger than that in the WT protein (21.6 s −1 vs. 16.4 s −1 at 14.10 T; 21.8 s −1 vs. 17.9 s −1 at 19.97 T). These alterations might be caused by the rotation of the Tyr128 side chain, as described above.
The G127V mutation also changed the R 2 rates of the residues located within the α3 helix. Because of the retroflexion of the Pro158 pyrrolidine, as described above, the R 2 rates of Val209 and Val210 subtly fluctuated at 14.10 T and were markedly disturbed at 19.97 T. Compared with the WT protein, the mutant showed slightly larger R 2 rates for the two residues at 14. Reduced spectral density mapping. To explicitly explore the internal motion of the amide backbone, we calculated the reduced spectral density functions at three frequencies, J(0), J(ω N ) and J(0.87ω H ), based on Figure 3. A comparison of the backbone dynamics parameters from HuPrP(G127V) and WT HuPrP derived from 15 N relaxation data. All NMR spectra were acquired at magnetic field strengths of 14.10 T (red for G127V, violet for WT) and 19.97 T (blue for G127V, olive for WT). experimentally derived 15 N relaxation data for both HuPrP(G127V) and WT HuPrP (Fig. 4). For HuPrP(G127V), the J(0) values of the N-terminal flexible segment and the C-terminus were less than 2.5 ns/rad for both magnetic fields. However, the C-terminal structural core displayed J(0) values varying from 5.0 ns/rad to 10.0 ns/rad. The α3 helix exhibited higher J(0) values than the α1 and α2 helices, but the α2-α3 loop displayed relatively smaller J(0) values than the α1 and α2 helices. Moreover, the N-terminal flexible segment showed J(ω N ) values scattering from 0.05 ns/rad to 0.35 ns/rad at the two magnetic fields, but the C-terminal structural core exhibited J(ω N ) values fluctuating near 0.27 ± 0.03 ns/rad at 14.10 T and 0.20 ± 0.03 ns/rad at 19.97 T. Furthermore, the N-terminal flexible segment showed J(0.87ω H ) values between 0.014 ns/rad and 0.045 ns/rad at 14.10 T, which changed to 0.011 ns/rad and 0.026 ns/rad at 19.97 T. The C-terminal structural core displayed J(0.87ω H ) values varying near 0.006 ns/rad at 14.10 T and 0.003 ns/rad at 19.97 T (Fig. 4). On the whole, compared with HuPrP(G127V), WT HuPrP did not show distinctly different J(0), J(ω N ), J(0.87ω H ) values or trends.
Compared with WT HuPrP, HuPrP(G127V) showed much larger J(0) values for the SS1-located residue Gly131 (14.5 ns/rad vs. 9.2 ns/rad at 14.10 T; 16.7 ns/rad vs. 12.5 ns/rad at 19.97 T). The J(0) values for Tyr163 located in the SS2 of the mutant were only subtly larger than those in the WT protein (8.1 ns/rad vs. 7.5 ns/rad at 14.10 T; 10.4 ns/rad vs. 9.5 ns/rad at 19.97 T), similar to the R 2 rate for Tyr163, which was slightly higher in the mutant than that in the WT protein. These results suggest that the two residues in the mutant underwent slow conformational fluctuations.

Relaxation dispersion measurements.
To compare in detail the dynamic features between HuPrP(G127V) and WT HuPrP, especially for Gly131 and Tyr163 with large J(0) values, we performed CPMG RD experiments at  Table S4.
For HuPrP(G127V), Gly131 and Tyr163 in the SS1 and SS2 segments displayed k ex rates of 1295 ± 122 s −1 and 2842 ± 186 s −1 , respectively (Fig. 5, Table S4). Notably, residue Tyr128 disappears from the 1 H-15 N HSQC spectrum because of peak broadening that may be caused by conformational exchange. The relaxation dispersion data from the Val161 and Tyr162 residues were not suitable for the CPMG RD analysis because of resonance overlapping. Interestingly, Met129 and Leu130 did not display observable conformational fluctuations (Fig. S9, Table S4). Gln172 and Gln186, located in the α2 helix, exhibited significant conformational exchanges, with k ex rates of 3171 ± 302 s −1 and 2143 ± 328 s −1 , respectively (Fig. 5, Table S4).
Additionally, the α3 helix in both the mutant and WT protein displayed substantial magnetic field strength-dependent conformational fluctuations on the μs-ms timescale (Fig. S10, Table S4). For example, two residues located in the α3 helix, Met205 and Thr216, exhibited significant conformational exchanges at 19.97 T rather than at 14.10 T.

Molecular dynamics simulations.
To further disclose the differences in dynamic structural properties between HuPrP(G127V) and WT HuPrP, we performed MD simulations based on the identified protein structures ( Fig. 6 and S11). By analysing secondary structure elements and the geometric relationship of residues Leu125-Asp167, we summarized five primary distinctions between the mutant and WT proteins: the SS1 and SS2 segments rarely form a β-sheet in the mutant, instead, two β-strands always formed a stable β-sheet in the WT during the entire MD simulation (Fig. 6a); the α1 helix in the mutant is extended compared with that in the WT protein (Fig. 6a); the Tyr128 side chain adopts either the "mediate" or "out" conformation in the mutant instead of the "in" conformation found in the WT protein (Fig. 6b,c); the dynamic distance between the mass centres of Val127 and Pro165 in the mutant is smaller than that of Gly127 and Pro165 in the WT protein (Fig. 6d); the dihedral angle psi(N-Cα-C-N) of Tyr157 is approximately 60° for the mutant but is approximately 180° in the WT protein (Fig. 6e). These distinct dynamic properties might derive from the difference in hydrophobicity between Val127 and Gly127. Compared with glycine, valine is more hydrophobic and tends to be near the hydrophobic Pro165 in the SS2-α2 loop, as supported by the 3D structures of the mutant and WT proteins. This spatial alteration might introduce alterations in the conformation of Tyr128 and Tyr157 and the feasibility of β-sheet formation.
More meaningfully, the G127V mutant induces intramolecular steric hindrance in the relatively larger Val127 side chain, leading to a striking structural rearrangement and conformational alternation of the Tyr128 side chain. The MD simulations suggest that the orientation of the Tyr128 side chain directly determined the feasibility of the intermolecular dimerization. For the "exposed" case (either the "mediate" conformation or "out" conformation) in HuPrP(G127V), the steric hindrance closely associated with the Tyr128 side chain potentially prevents the monomeric prion protein from forming intermolecular interactions and might thus prohibit prion dimerization. In contrast, for the "buried" case (the "in" conformation) in WT HuPrP, the Tyr128 side chain likely does not reduce the feasibility of intermolecular dimerization (Fig. 6c). Hence, the mutation-induced structural rearrangement and dramatic conformational exchange of the Tyr128 side chain might be unfavourable for the dimerization and conformational conversion of HuPrP(G127V).

Discussion
Prion disease pathogenesis is closely associated with the conformational conversion of prion proteins from PrP C to PrP Sc . The α2 and α3 helices 33,53,54 , octarepeats 55 , the N-terminal flexible segment 56 , and the glycophosphatidylinositol (GPI) anchor 57 contained in PrPs might be involved in conformational conversions [58][59][60][61] . Moreover, conformational conversion is triggered at the two β-strands, the α1 helix, the α2 helix, the β1-α1 loop, the α1-β2 loop, and the β2-α2 loop 34,35,[62][63][64][65][66][67] . Notably, the more stable β-structure is formed by the segment of the N-terminus (residues 120-144), the earlier stages of misfolding are caused by 43,68,69 in MD, and a relatively short β-sheet core (residues 112-139) is capable of seeding the conversion to fibrils in vitro 70 . Nevertheless, the molecular mechanism underlying the disease-resistant effect of the G127V mutation still remains elusive. To reveal the molecular mechanisms, we determined the solution structures of both the HuPrP(G127V) and WT HuPrP under identical experimental conditions. We then analysed the backbone dynamics using 15 N relaxation experiments and conducted MD simulations for both proteins. We focused primarily on the dynamic structural properties of the two SS segments and adjacent regions, including intramolecular interactions between SS1 and SS2, SS1/SS2 and α2, SS1/SS2 and α3, α1/α1-SS2 loop and α3, and SS2-α2 loop/α2 and α3.
The primary structural distinction between HuPrP(G127V) and WT HuPrP 22,33 or other pathogenic mutants 16,23,[25][26][27][28] is that HuPrP(G127V) extends atomic distances between SS1 and SS2, increases the solvent accessibility surface of SS1-located residues (Figs S5 and S6), and exhibits significant μs-ms timescale conformational fluctuations at Tyr128, Gly131 and Tyr163. These properties indicate that the SS region is more flexible than the β-sheet and is not prone to conversion to a stable β-sheet conformation. Moreover, the striking structural rearrangement and alternate conformation of the Tyr128 side chain potentially induces the intermolecular steric hindrance effect, prevents the formation of intermolecular hydrogen bonds and prohibits prion protein dimerization. Notably, our result is fundamentally different from a previously published result, which suggested that the intermolecular steric hindrance was closely associated with the bulky sidechain of Val127 71 . The previous MD simulation work was based on the modelled structures of the G127V mutant using the solution structure of WT HuPrP (125-228) determined at pH 7.0 (PDB ID: 1HJN) and the crystal structure of the β1-strand fragment (PDB ID: 4TUT) as the templates 71 . Additionally, HuPrP(G127V) also alters the local electrostatic potential distribution near the SS1 and SS2 segments to influence potentially electrostatic interactions.
Previous studies suggest that pathogenic and protective mutants of PrPs have similar structures and dynamics 29,30,35,36,42,44 . However, our results confirmed that the structural and dynamic alterations caused by G127V are tremendously different from the changes caused by Met129, Val129 or any other known mutants 29,30,35,36,42,44 . Furthermore, as previously hypothesized, the β-sheet in the prion protein, and especially the β1-strand, might be the cornerstone on which prion protein aggregation is triggered [34][35][36]63,64,66 . For instance, D178N/M129 and F198S form intermolecular antiparallel four-strand β-sheets based on β1-strands in crystal structures 34 , and the β1-strand fragments form a steric zipper conformation 35,36 . However, HuPrP(G127V) possesses flexible SSs with structural rearrangement and conformational fluctuations, rearrangement and alternate conformation of the Tyr128 side chain as well as surface electrostatic potential redistribution that destroys the prion protein aggregation trigger and prohibits prion protein fibrillization.
On the other hand, HuPrP(G127V) and WT HuPrP have similar atomic distances between Met129 and Tyr163 (Table S2), similar H/D exchanges of Met129, Leu130, Val161, Tyr162 and Tyr163 (Figs S7 and S8), and similar dynamic properties for Met129 and Leu130 (Figs 3, 4 and S9). The similar structural and dynamic properties between the SSs in the mutant protein and the β-sheet in the WT protein imply that HuPrP(G127V) might partially reserve the structural and dynamic properties of the β-sheet in the WT protein via the SS pattern.
Regarding the intramolecular interactions between the SSs and the α2/α3 helices in the G127V mutant, we found that (I) the G127V mutant changes the orientation of the Tyr128 side chain and leads to different conformational exchanges for Ile182 and Gln186 (Table S4); (II) the mutation-induced steric hindrance effect between the side chains of Val127 and Arg164 pushes the Arg164 side chain close to Asp178, strengthens the electrostatic interaction between Arg164 and Asp178 (Fig. S2e,f) and enhances the hydrophobic interaction between Val127 and Pro165 (Fig. 6d); (III) the G127V mutant positions two SS2-located residues, Tyr163 and Arg164, slightly closer to Cys179 (Fig. S2g,h). These structural alternations reveal that the G127V mutation changes the local circumstances around the SSs and α2/α3 regions in HuPrP(G127V), which are distinctly different from those in the WT protein and several other HuPrP mutants such as D178N 34 . These unique structural features of HuPrP(G127V) potentially reduce the feasibility of prion protein aggregation 54 .
Furthermore, distinguishing structural features are also identified in the regions around the α1 helix and the α1-SS2 loop and the α2 and α3 helices in HuPrP(G127V), which may be responsible for the prion disease-resistance effects of the G127V mutant. Overall, the G127V mutation extends the α1 helix and induces the retroflexion of the Pro158 pyrrolidine, thus increasing the curvature of the α1-SS2 loop. These structural alterations potentially prevent the unwinding of the α1 helix. As previously suggested, the α1 helix could be converted to the β-strand to form fibrils via a despiralization process 63,64 . In the G127V mutant, the atomic distances between Tyr157, Pro158, Val209 and Val210 changed (Fig. 2g,h) and introduced magnetic field strength-dependent fluctuations in the R 2 and R 2 /R 1 rates of Val209 and Val210 (Fig. 3). These alterations might correspondingly change the local environment of the Val210 mutable site (the V210I mutant is associated with fCJD 27,49 ), and may promote the protective effect of the G127V mutant.
Additionally, the extended α1 helix, the bent α2 helix, and the α3 helix are packed more compactly in HuPrP(G127V) than those in the WT protein. The unique geometric packing in the G127V mutant is similar to the protective packing in the V209M mutant 16 and might slow the initial fibrillization rate in a manner similar to that in HuPrP(V209M) 16 and the G126V mutant of the mouse prion protein (moPrP) 72 . The moPrP(G126V) is equivalent to HuPrP(G127V), slows initial fibril growth and increases the critical concentration 72 . The compact geometric packing might also change the local environment of the α2-α3 loop near the α1 helix. Note that the fCJD-associated F198S 34  potential distribution on the region encompassing the α1 and α3 helices is diametrically distinct from those in the WT and the fCJD-associated E200K mutant 26 . The alterations of HuPrP(G127V) in both the geometric packing and electrostatic potential distribution combined with the close atomic distances between SS2 and the disulfide bridge, might prohibit rearrangement of the disulfide bridge, aggregation and fibrillization as previously published results 16,33,54,72 .
Compared with the WT HuPrP, the SS2-α2 loop (Pro165-Asn171) of the HuPrP(G127V) exhibits more flexibility. The G127V mutation allows Met166 at the SS2-α2 loop to be closer to Tyr218, which is located in the α3 helix (Fig. S2i,j). In HuPrP(G127V), Gln172, next to the SS2-α2 loop, undergoes a more significant conformational exchange than that in the WT HuPrP (Table S4). These results indicate that the SS2-α2 loop has dynamic structural features distinct from the β2-α2 loop (Pro165-Gln172), which is probably correlated to the susceptibility to prion disease 65 . The unique dynamic structural properties of the SS2-α2 loop might contribute to the prion disease resistance of the G127V mutant as well.
Astonishingly, the α3 helix in HuPrP(G127V) showed R 2 /R 1 ratios that were dramatically different from those of the WT HuPrP at 19.97 T (Fig. 3). Moreover, the α3 helix exhibited varying J(0) values, similar to the R 2 /R 1 ratios. In addition, Met205 and Thr216 in both proteins experienced slow conformational exchange, which was observable only at 19.97 T. Unexpectedly, the α3-located Glu219 in both proteins displayed large R 2 rates and J(0) values but did not exhibit observable conformational exchanges (Fig. S10, Table S4). Furthermore, Glu219 in HuPrP(G127V) showed the dynamic property, distinct from HuPrP(E219K) 23 . Thus, the dramatically altered dynamic structural properties relevant to the α3 helix could potentially influence the intermolecular interactions of the prion protein with the so called "protein X" 73,74 .
Besides, our fibrillization experiments showed that HuPrP(G127V) had significantly slower initial fibril growth than WT HuPrP. The measured lag phases were 61 ± 2 h for HuPrP(G127V) and 25 ± 2 h for WT HuPrP as showed in Fig. S12. Moreover, the mixing samples of WT HuPrP and HuPrP(G127V) (at a mixing ratio of 1:1) exhibited a slower fibrillization rate than WT HuPrP but faster than HuPrP(G127V). The measured lag phase was 47 ± 2 h for the mixing sample. These kinetic analyses are similar to the quantitative comparison of moPrP(G126V) and WT moPrP 72 . These unique dynamic structural features might be responsible for the prion disease-resistance effect of the G127V mutant 20,21 . As expected, the further study of the exploitation of the structural and dynamic features of the GSS-associated mutant G131V 75,76 (GSS), which was confirmed to enhance the stability of the β-sheet and drive conformational conversion by MD simulation 77,78 , would greatly help to address the crucial role of the SS1 segment in conformational conversion and propagation.
Summarily, we performed solution structure determinations, NMR dynamics analysis and MD simulations on both HuPrP(G127V) and WT HuPrP. We addressed the G127V mutation-induced significant distinct alterations in structural and dynamic properties in detail. The G127V mutation extends atomic distances between the SS1 and SS2 segments and enhances the conformational exchange of the two strands, leading to the formation of the SS pattern instead of the stable β-sheet. The relatively larger hydrophobic side chain of Val127 introduces steric hindrance and a striking structural rearrangement in the Tyr128 side chain. Additionally, the G127V mutation also subtly alters the geometric stacking of the three α helices. These structural and dynamic features might prevent the SS1 (Tyr128-Gly131) and SS2 (Val161-Arg164) segments and adjacent regions from being converted into a stable β-sheet under certain circumstances. Furthermore, the steric hindrance effect of the rearrangement of the Tyr128 side chain, together with the dramatic conformational alternation, could potentially prohibit the prion protein intermolecular interaction and dimerization, and thus inhibit prion protein aggregation and fibrillization. Moreover, HuPrP(G127V) had significantly slower initial fibril growth than WT HuPrP. Although more researches are required to clarify completely the molecular mechanisms of the prion disease-resistance of HuPrP(G127V), our results provide several important evidences regarding the differences in structure and dynamics between HuPrP(G127V) and WT HuPrP. These structural and dynamic differences substantially contribute to the different conversion of monomer to dimer in MD and of monomer to fibril in fibrillization between the two proteins. This work may be helpful for mechanistically understanding the pathogenesis of prion diseases and for developing effective drugs against prion diseases.

NMR sample preparation.
Recombination of the pET30a plasmids without any tag bearing the DNA of the WT HuPrP (residues 91-231 with G127M129) was prepared as previously described [79][80][81] . The recombination plasmids for HuPrP(G127V) (residues 91-231 with the genotype of V127M129) were cloned by PCR using site-directed mutagenesis. The forward primer used in the PCR was: 5′-AGTGGTGGGGGGCCTTGGCGTTTACATGCTGGGAA-3′ and the reverse primer used was: 5′-ATGGCACTTCCCAGCATGTAAACGCCAAGGCCCCCCA-3′. The uniformly labelled protein was overexpressed in E. coli Bl21(DE3) grown in M9 medium. 15 NH 4 Cl was added to the M9 medium to prepare 15 N-labelled proteins and both 15 NH 4 Cl and 13 C 6 -glucose were added to prepare 13 C/ 15 N-labelled proteins. After the cells were sonicated and the lysates were centrifuged, the inclusion bodies were denatured in 6 M guanidine hydrochloride and refolded by dialysing against the NMR buffer (20 mM NaOAc, 0.02% NaN 3 , pH 4.5) as previously described [79][80][81] . Thereafter, the protein was purified in NMR buffer through size exclusion chromatography with Superdex-75 on an ÄKTA FPLC system (GE Healthcare). Finally, the protein solution was concentrated to approximately 0.5 mM with 10% D 2 O (v/v). time of 120 ms was used for both 15 N-edited NOSEY-HSQC and 13 C-edited NOESY-HSQC experiments. All NMR spectra were processed by NMRPipe software 82 and analysed with CARA software 83 .
Structure calculations. Distance constraints were generated from the 1 H-1 H NOEs of both 13 C and 15 N-labelled NOESY-HSQC spectra. Dihedral angle restraints were obtained based on chemical shifts of the backbone atoms including HN, Hα, Cα, Cβ, C(O), and N using the TALOS+ programme 84 . The 3D structures were calculated and refined with the XPLOR-NIH package 85 . Then, the qualities of the calculated structures were evaluated by the PROCHECK programme 86  The repeated spectra were used for experimental error analysis. { 1 H}-15 N NOEs were obtained by recording spectra with a 1 H pre-saturation of 3 s plus a 2-s relaxation delay and without a pre-saturation of a 5-s relaxation delay. All NMR spectra were processed using NMRPipe software 82 and analysed using CcpNmr software 89  Reduced spectral density mapping. Reduced spectral density mapping is usually employed to characterize the internal motions of the N-H bonds with the assumption that J(0.87ω H ) is approaching J(ω H + ω N ) and J(ω H + ω N ) at high frequencies 90 . Therefore, the values of the relaxation rates R 1 , R 2 and { 1 H}-15 N NOEs are taken to map the spectral density using the following formula:

Relaxation dispersion measurements.
Single quantum CPMG RD experiments were performed on the same NMR instruments described above (850 MHz at 19.97 T with a TCI cryogenic probe, 600 MHz at 14.10 T with a BBO cryogenic probe). The CPMG RD spectra were recorded on 15 N-edited HuPrP(G127V) and WT HuPrP proteins at 25 °C and pH 4.5 using a constant relaxation time of 40 ms and under thirteen ν CPMG values of 0, 100(×2), 200, 300, 400, 500, 600, 700(×2), 800, 900, and 1000 Hz. All spectra were recorded with complex points of 1024 × 128. The ν CPMG is defined by the following formula 92 : Here, τ cp is the time between refocusing pulses during the CPMG pulse train. We used the following equation 92 where T cp is the constant transverse relaxation time and I(ν CPMG ) and I 0 are the intensity with or without different ν CPMG . The RD of R eff 2 relies on ν CPMG if the residue undergoes conformational exchange at the μs-ms timescale. All spectra were processed in NMRPipe 82 and the integrals of the peaks were obtained in NMRFAM-Sparky 93 . The dispersion data were fitted with a Carver-Richards two-state exchange model 94 in NESSY software 95 . Similar to the 15 N backbone dynamics analysis, overlapping amide resonances were not used for CPMG RD analysis.

Molecular Dynamics
Simulations. All MD simulations were performed with the AMBER99SB 96 force field in AMBER12 97 . All systems were solvated within a cubic box of TIP3P 98 water molecules by extending 10 Å from the protein surface. The initial coordinates and topology files were generated using the tleap programme contained in AMBER12. First, energy minimizations were performed to relax the solvent and optimize the system. Then, each system was gradually heated from 0 to 300 K under the NVT ensemble for 100 ps and another 100 ps of NPT ensemble MD simulation was performed at 300 K and a target pressure of 1.0 atm. Finally, a 100 ns MD simulation under the NVT ensemble was performed for each model. The system temperature was controlled by the Langevin thermostat method. During the MD simulations, all hydrogen-containing bonds were constrained using the SHAKE algorithm 99 . A cut-off of 12 Å was set for both the van der Waals and electrostatic interactions. The DSSP algorithm was employed to assign the secondary structure of the protein 100 .
Residual Dipolar Couplings. Initially, both 15 N-labeled HuPrP(G127V) and WT HuPrP were dissolved in H 2 O buffer (90% H 2 O, 10% D 2 O, 20 mM NaOAc, 0.02% NaN 3 , pH 4.5) to a final concentration of 0.4 mM. As reference spectra, 2D 1 H-15 N IPAP-HSQC spectra were recorded at 25 °C on a Bruker Avance III 600-MHz spectrometer (magnetic field strength of 14.10 T with a triple-resonance TCI cryogenic probe) at the University of Science and Technology of China. All spectra were recorded with complex points of 1024 × 400. Then, the two proteins were dissolved in C 12 E 5 /n-hexanol alignment media 101 . The final concentration of C 12 E 5 was 3% (r = 0.96) 101 . 2D 1 H-15 N IPAP-HSQC spectra were recorded under the same experimental conditions. All data were processed on NMRPipe 82 , analysed on NMRFAM-Sparky 93 and fitted on PALES 102 . The Q-value was fitted by PALES, which is normally used to assess the agreement between the experimental RDCs and calculated RDCs based on the structure 102,103 . When fitted using the PALES program, the experimental RDCs were just from the residues of the C-terminal structural core minus the overlapping resonance, as described above.
Amide Hydrogen/Deuterium Exchange. Both 15 N labelled HuPrP(G127V) and WT HuPrP were initially dissolved in H 2 O buffer (90% H 2 O, 10% D 2 O, 20 mM NaOAc, 0.02% NaN 3 , pH 4.5). As reference spectra, 2D Fast-1 H-15 N HSQC 104 were recorded at 25 °C on a Bruker Avance III 850-MHz spectrometer (magnetic field strength of 19.97 T with a triple-resonance TCI cryogenic probe). All spectra were recorded with complex points of 1024 × 128. Through buffer exchange with centrifugal filter devices (Amicon ® Ultra 3 K device) at 2,555 × g and 4 °C for 3 h, the proteins were re-dissolved in equal volumes of D 2 O buffer (99.9% D 2 O, 20 mM NaOAc, 0.02% NaN 3 , pH 4.5). Then, 2D Fast-1 H-15 N HSQC spectra were recorded on the re-dissolved proteins as amide proton exchange spectra under the same experimental conditions. All data were processed on Topspin 3.2 (Bruker) and analysed in CcpNmr 89 . This approach allowed the quantitative analysis of peak intensity decreases caused by the mutation but could not be used to measure amide protection factors for the protein 105 .
Accession codes. Chemical shift data were deposited in the Biological Magnetic Resonance Data Bank (http://www.bmrb.wisc.edu) under accession numbers 27259 for HuPrP(G127V) and 27264 for WT HuPrP. The atomic coordinates were deposited in the Protein Data Bank under the accession codes 5YJ4 for HuPrP(G127V) and 5YJ5 for WT HuPrP.