The conformational change from the cellular prion protein (PrPc) to scrapie prion protein (PrPsc) is a key process in prion diseases. The prion protein has buried water molecules which significantly contribute to the stability of the protein; however, there has been no report investigating the influence on the buried hydration sites by a pathogenic mutation not adjacent to the buried hydration sites. Here, we perform molecular dynamics simulations of wild type (WT) PrPc and pathogenic point mutant T188R to investigate conformational changes and the buried hydration sites. In WT-PrPc, four buried hydration sites are identified by residence time and rotational relaxation analysis. However, there are no stable buried hydration sites in one of T188R simulations, which indicates that T188R sometimes makes the buried hydration sites fragile. We also find that fluctuations of subdomains S1-H1-S2 and H1-H2 increase in T188R when the buried hydration sites become unstable. Since the side chain of arginine which is replaced from threonine in T188R is larger than of threonine, the side chain cannot be embedded in the protein, which is one of the causes of the instability of subdomains. These results show correlations between the buried hydration sites and the mutation which is far from them, and provide a possible explanation for the instability by mutation.

Prion diseases, including Creutzfeldt-Jakob disease and mad cow disease, are one of the neurodegenerative diseases caused by aggregation of misfolding proteins.1 The prion protein is a causal protein for prion diseases and its dynamics has attracted researchers’ interests for elucidating prion diseases. The prion diseases are composed of mainly three processes: (i) conformational change of secondary structure, (ii) aggregation of transformed prion proteins, and (iii) accumulation of protein amyloids on nerve cells. Conformational change and aggregation of prion proteins are quite important because they may be related to the mechanism of Alzheimer’s disease and Parkinson disease.2,3 However, these processes remain an unresolved biological phenomena.

The prion protein has two small β-sheets (β1, β2) located near C-terminal and three α-helices (H1, H2, H3) with a disulfide bond between C179 and C214.4 The secondary structural change from the α-helix-rich cellular prion protein (PrPc) to a protease-K resistant β-sheet-rich scrapie prion protein (PrPsc) is also a key process of prion diseases,5 and is caused by low pH,6–9 high temperature,10 or pathogenic mutations.11–15 

Water molecules play an important role for biochemical properties,16 e.g. protein conformation,17 aggregation of proteins, docking of proteins,18,19 interaction with other biological materials,20 and the stability of biological materials.21–23 Water molecules in hydration shell are usually located around a protein; however, several water molecules are buried in the protein, called buried water molecules.24–26 Buried water molecules have hydrogen bonds with amino acids, and contribute to the formation of secondary structure and the stability of protein.27 2 Å crystal structure has showed that prion protein also has buried water molecules.28 Using all-atom molecular dynamics (MD) simulations, De Simone et al. clarified the location of the buried water molecules and showed that these buried water molecules are quite important for the stability of secondary structure.29 They have also investigated an influence on the buried hydration sites by a pathogenic point mutation (Q217R) located nearby the buried hydration sites. The results showed that the pathogenic point mutation disrupts the buried hydration site. However, some recognized pathogenic mutations are far from the sites. Therefore, it is important to investigate an influence on the buried hydration sites because there are no studies so far clarifying the effect.

In this article, we perform MD simulations of both the human prion protein wild type (WT)-PrPc and the pathogenic mutant prion protein (T188R) to clarify an influence of the pathogenic point mutation T188R on the buried hydration sites. The T188R mutation, which is not adjacent to the buried hydration sites, has been investigated by both experiments30,31 and MD simulation.32 Our results provide a first evidence that the pathogenic point mutation affects the stability of buried water molecules far from the mutation point.

We performed two independent all-atom MD simulations for two systems: the human WT-PrPc and the pathogenic mutant T188R. Every simulation was performed totally for 1.2 μs. The initial structure of WT-PrPc was obtained from NMR structural data (Protein data band code:1HJM) which contains residues 125-228 with a disulfide bond between C179 and C214.33 We made an initial structure of T188R using the Swiss-PdbViewer34 by replacing residue 188 from threonine to arginine. Each prion protein was put on the center of a cuboid box and solvated with TIP3P water molecules. Systems contained 12,761 water molecules and 3 and 2 sodium ions in WT-PrPc and T188R as counterions, respectively. All MD simulations were carried out with NAMD 2.9 simulation package.35 The AMBER ff99SBildn force field was used.36 The SHAKE algorithm was used for restraint of the covalent bonds of both prion proteins and water molecules. Langevin dynamics thermostat and Nosé-Hoover Langevin piston barostat were applied for the temperature and pressure coupling. Switching cutoffs starting from 0.8 nm to 1.4 nm was applied to both electrostatic and van der Waals interactions. The Particle Mesh Ewald method was used to calculate the electrostatic interactions. The periodic boundary condition was applied for each system. First, we performed 1000 steps of minimization using the conjugate gradient algorithm only for water molecules. After isothermal 1 ns NVT simulation (number of particles, volume, and temperature) at a temperature of 310 K, we performed two 600 ns isothermal isobaric NPT simulations (number of particles, pressure and temperature) for each system at 310 K and a pressure of 0.1 MPa with a time step of 2 fs.

Cα-root-mean-square deviation (Cα-RMSD) is an indicator of stability and conformational change of proteins. Since we are only interested in the region related to secondary structure (from residue 125 to 225), three N and C terminal residues are omitted from following analysis. Figure 1(A) shows the Cα-RMSDs of WT-PrPc and T188R for all simulations. All Cα-RMSDs reached equilibrium states at 200 ns; therefore, we determined that every system reached an equilibrium state at 200 ns. Later analysis were conducted using the data after equilibrium, which accounts for totally 800 ns simulation data for WT-PrPc and T188R.

FIG. 1.

(A) The Cα-RMSD from 0 ns to 500 ns. We used the notations 1 and 2 to express two different simulatons for WT-PrPc and T188R. The reference structure is crystal structure obtained from Protein Data Bank. We have calculated the Cα-RMSDs after fitting to the reference. (B) Cα-RMSF of WT-PrPc and (C) Cα-RMSF of T188R averaged from 200 ns to 600 ns. The vertical axis shows amino acid number. Black arrow indicates mutation point residue 188.

FIG. 1.

(A) The Cα-RMSD from 0 ns to 500 ns. We used the notations 1 and 2 to express two different simulatons for WT-PrPc and T188R. The reference structure is crystal structure obtained from Protein Data Bank. We have calculated the Cα-RMSDs after fitting to the reference. (B) Cα-RMSF of WT-PrPc and (C) Cα-RMSF of T188R averaged from 200 ns to 600 ns. The vertical axis shows amino acid number. Black arrow indicates mutation point residue 188.

Close modal

For analysis of structural flexibility, we calculated the Cα-root mean square fluctuation (Cα-RMSF) for all simulations. The Cα-RMSF indicates the intensity of fluctuations, which is defined as

RMSF i = 1 T t = 1 T ( x i ( t ) x ¯ i ) 2 ,
(1)

where xi(t) is the coordinate of Cα of the ith amino acid, x i ̄ is the averaged coordinate and T is the measurement time after reaching equilibrium state. Before the RMSF analysis, we rotate the structure to minimize the Cα-RMSD. Figures 1(B) and 1(C) show the Cα-RMSFs of WT-PrPc and T188R. The distributions of the Cα-RMSF values are not dramatically different. There are five peaks in the RMSFs of WT-PrPc and T188R. These results show that secondary structure does not dramatically change by the point mutation in this time scale.

The secondary structure is defined by the Dictionary Secondary Structure of Protein (DSSP) method.37 Figures 2(A) and 2(B) show the time series of secondary structure and snapshots of the WT-PrPc and the T188R, respectively. We confirm that there are two stable sheets and three helical structures in all simulations. In T188R, the secondary structure is stable after around 200 ns. Secondary structure does not dramatically change by the mutation; however, several regions change into different secondary structures. By the T188R mutation, a part of helix 2 (residue 174 to 178) disrupts and changes into a coil structure, which could induce the transition into the scrapie form,38–40 and β-sheet 2 also increases. β-sheets are putative elements for aggregation.41 Table I shows the proportion of secondary structures. This also indicates that T188R causes a decrease of α-helix content and an increase of β-sheet content.

FIG. 2.

Secondary structure of WT-PrPc and T188R. (A) Evolution of secondary structure of WT-PrPc (upper) and T188R (bottom). Secondary structure is calculated by DSSP method. C, T, G, H, I and B indicate coil, turn, 310-helix, α-helix, π-helix and β-sheet or anti-parallel β-sheet, respectively. (B) The secondary structures of WT-PrPc and T188R. The secondary structure is the mean coordinates. The α-helix, β-sheet, random coil, and turn regions are colored orange, yellow, green, and gold. Residue 188 is shown as beads form.

FIG. 2.

Secondary structure of WT-PrPc and T188R. (A) Evolution of secondary structure of WT-PrPc (upper) and T188R (bottom). Secondary structure is calculated by DSSP method. C, T, G, H, I and B indicate coil, turn, 310-helix, α-helix, π-helix and β-sheet or anti-parallel β-sheet, respectively. (B) The secondary structures of WT-PrPc and T188R. The secondary structure is the mean coordinates. The α-helix, β-sheet, random coil, and turn regions are colored orange, yellow, green, and gold. Residue 188 is shown as beads form.

Close modal
TABLE I.

Proportion of secondary structures in WT-PrPc and T188R with standard errors.

α-helix 310-helix pi-helix β-sheet anti β-sheet
WT-PrPc 1  49.17 ± 0.019  6.14 ± 0.017  0.003 ± 0.000  2.72 ± 0.003  8.10 ± 0.007 
WT-PrPc 2  50.47 ± 0.018  5.29 ± 0.048  0.000 ± 0.000  3.85 ± 0.000  5.77 ± 0.000 
T188R 1  46.44 ± 0.018  7.69 ± 0.021  0.003 ± 0.000  1.92 ± 0.000  11.54 ± 0.000 
T188R 2  46.35 ± 0.020  8.60 ± 0.022  0.000 ± 0.000  4.62 ± 0.004  9.26 ± 0.008 
α-helix 310-helix pi-helix β-sheet anti β-sheet
WT-PrPc 1  49.17 ± 0.019  6.14 ± 0.017  0.003 ± 0.000  2.72 ± 0.003  8.10 ± 0.007 
WT-PrPc 2  50.47 ± 0.018  5.29 ± 0.048  0.000 ± 0.000  3.85 ± 0.000  5.77 ± 0.000 
T188R 1  46.44 ± 0.018  7.69 ± 0.021  0.003 ± 0.000  1.92 ± 0.000  11.54 ± 0.000 
T188R 2  46.35 ± 0.020  8.60 ± 0.022  0.000 ± 0.000  4.62 ± 0.004  9.26 ± 0.008 

Hydration structure is directly related to conformation, thermodynamic properties, and stability of protein.21 Here, we focus on buried hydration sites. To identify the buried hydration sites, we used the mean residence time analysis. The residence time is defined as the time in which a water molecule resides in a sphere centered at an atom in the protein. We used amino-acid-based residence time, while previous study used grid-based residence time. Because secondary structure is not stable during the simulations, the amino-acid-based analysis is better than the grid-based analysis especially for the case of the mutant.29 The radius of the sphere was set to 0.4 nm which is larger than ordinal hydrogen bond length. However, this definition cannot distinguish whether a water molecule completely leaves from the site or not. Therefore, we determine that a water molecule leave from the site when the distance between the protein atom and the water molecule is greater than 0.8 nm.

In Figure 3, the mean residence times on each atom are mapped on protein atoms. Most of the mean residence times are less than 100 ps; however, there are quite high mean residence times at some atoms, i.e., the buried hydration sites [see table II]. From the high residence time areas, we identified four buried hydration sites, Site 1(G131, V161, and Q217), Site 2 (Q217, V220, and R228), Site 3 (F175, R163, and R164), and Site 4 (Y162, F175, and T183) in both WT-PrPc and T188R. Hence, the mutation does not disrupt the buried hydration sites.

FIG. 3.

The mean residence times were mapped on atoms of prion protein for each calculation. The buried water molecules are explicitly visualized in the sites with residue 188 shown as beads form.

FIG. 3.

The mean residence times were mapped on atoms of prion protein for each calculation. The buried water molecules are explicitly visualized in the sites with residue 188 shown as beads form.

Close modal
TABLE II.

Mean residence times, total residence times of A WT-PrPc and B) T188R.

(A) WT-PrPc
PrPc 1 Mean [ns] Total [ns] PrPc 2 Mean [ns] Total [ns]
Site 1  199.8 ± 37.81  399.6  Site 1  3.577 ± 2.509  114.5 
Site 2  57.34 ± 52.87  401.4  Site 2  7.413 ± 4.383  244.6 
Site 3  62.94 ± 33.93  377.9  Site 3  10.94 ± 5.469  404.7 
Site 4  7.277 ± 2.890  356.7  Site 4  53.22 ± 28.57  266.1 
(B) T188R 
T188R 1  Mean [ns]  Total [ns]  T188R 2  Mean [ns]  Total [ns] 
Site 1  0.202 ± 0.002  31.28  Site 1  13.08 ± 8.207  366.2 
Site 2  3.182 ± 0.736  308.6  Site 2  9.159 ± 5.139  338.9 
Site 3  2.955 ± 0.526  428.5  Site 3  15.13 ± 8.914  408.0 
Site 4  6.187 ± 1.043  377.4  Site 4  42.78 ± 22.24  299.4 
(A) WT-PrPc
PrPc 1 Mean [ns] Total [ns] PrPc 2 Mean [ns] Total [ns]
Site 1  199.8 ± 37.81  399.6  Site 1  3.577 ± 2.509  114.5 
Site 2  57.34 ± 52.87  401.4  Site 2  7.413 ± 4.383  244.6 
Site 3  62.94 ± 33.93  377.9  Site 3  10.94 ± 5.469  404.7 
Site 4  7.277 ± 2.890  356.7  Site 4  53.22 ± 28.57  266.1 
(B) T188R 
T188R 1  Mean [ns]  Total [ns]  T188R 2  Mean [ns]  Total [ns] 
Site 1  0.202 ± 0.002  31.28  Site 1  13.08 ± 8.207  366.2 
Site 2  3.182 ± 0.736  308.6  Site 2  9.159 ± 5.139  338.9 
Site 3  2.955 ± 0.526  428.5  Site 3  15.13 ± 8.914  408.0 
Site 4  6.187 ± 1.043  377.4  Site 4  42.78 ± 22.24  299.4 

Figure 4(A) shows the locations of the four buried hydration sites superposed on the secondary structure of WT-PrPc 1. All buried water molecules are basically fixed by three amino acids. Sites 1 and 2 are composed of G131, V161, Q217, R220, and R228 [see Fig. 4(B)]. In Site 1, the buried water molecule has hydrogen bonds with G131 and V161 as a acceptor and Q217 as a donor. The buried water molecule in Site 2 also makes hydrogen bonds with Q217 as a donor and with R220 and R228 as a acceptor. In other words, Q217 is structurally stabilized by two buried water molecules.

FIG. 4.

Position of the water molecules in WT-PrPc 2. (A) The location of the buried water molecules superposed onto the secondary structure of WT-PrPc. Amino acids forming hydrogen bonds with the buried water molecules are explicitly visualized. The pathological mutation is shown as beads shape. (B) Hydrogen bond networks with amino acids around the buried water molecules. Top and bottom buried hydration sites are site 1 and 2. (C) Hydrogen bond networks around site 3 (top) and 4 (bottom).

FIG. 4.

Position of the water molecules in WT-PrPc 2. (A) The location of the buried water molecules superposed onto the secondary structure of WT-PrPc. Amino acids forming hydrogen bonds with the buried water molecules are explicitly visualized. The pathological mutation is shown as beads shape. (B) Hydrogen bond networks with amino acids around the buried water molecules. Top and bottom buried hydration sites are site 1 and 2. (C) Hydrogen bond networks around site 3 (top) and 4 (bottom).

Close modal

The buried water molecules in Sites 3 and 4 are wrapped by Y162, Y163, R164, F175, C179, and T183 [see Fig. 4(C)]. In Site 3, the buried water molecule is fixed by hydrogen bond with F175 as a donor and with R164 or Y163 as a acceptor. Site 4 is composed of Y162 and F175, and T183 as an acceptor and a donor, respectively. These buried hydration sites are stable during the MD simulations. Table II shows that the mean residence times of the buried hydration sites are much longer than those of non buried hydration sites. We confirmed that Sites 1 and 3 are identical to two of the buried hydration sites reported in De Simone et al.29 Site 4 was not reported in previous study, but Sites 4 is located quite near Site 3. In other words, grid-based analysis cannot distinguish these sites. Site 2 was also not reported because R228 was truncated in previous study.

We also identify the four buried hydration sites in T188R. However, in T188R 1, the mean residence times become smaller, and there are no areas which have long mean residence times (more than 10 ns). This result implies that the buried hydration sites become unstable, and the sites cannot keep hydration configuration. On the other hand, the several buried hydration sites are still stable in T188R 2. These results show that the pathogenic point mutation T188R sometimes makes the buried hydration sites unstable.

To investigate a mechanism of decrease of the mean residence times, we consider the total residence time defined as a sum of residence times of all water molecules residing at the site. There are two mechanisms of the decrease of mean residence times. First, the buried hydration site is more exposed to other bulk water molecules. In this case, the mean residence time decreases but total residence time increases or does not change. Second, the buried hydration site is more wrapped by amino acids and less exposed to other bulk water molecules. Water molecules trying to access to the site; however, they are interrupted by other amino acids, which leads to the decrease of both the total residence time and the mean residence time. In T188R 1, the mean residence times of Sites 2, 3 and 4 decrease while the total residence times do not change, which is caused by the first mechanism. However, both the total and the mean residence times of Site 1 decrease because of the second mechanism.

Because these buried water molecules are stabilized by hydrogen bonds, the rotational motions of the water molecules are also restricted; therefore, reorientational relaxation time becomes slow.42 We calculate the reorientational time correlation function according to the following equation:

C ( t ) = 1 2 3 u ( t 0 ) u ( t 0 + t ) 2 1 ,
(2)

where t0 is the entry time of a water molecule into the sphere, u(t0 + t) is the dipole vector of the water molecule at time t + t0. Figure 5 shows the reorientational time correlation function of buried water molecules within 1 ns. The results clearly show that the rotational relaxation of all buried water molecules is dramatically slower than that of bulk water molecules. All buried water molecules are fixed by hydrogen bonds with amino acids in all systems.

FIG. 5.

The reorientational correlation function of the buried water molecules in WT-PrPc 2. Blue, orange, red, green, and black lines show water molecules in Site 1, 2, 3, 4, and bulk, respectively. The inset is zoomed range from 0.5 to 1.0.

FIG. 5.

The reorientational correlation function of the buried water molecules in WT-PrPc 2. Blue, orange, red, green, and black lines show water molecules in Site 1, 2, 3, 4, and bulk, respectively. The inset is zoomed range from 0.5 to 1.0.

Close modal

Basically, the rotational relaxation of water molecules consists of two processes. First, hydrogen of water molecules librates around the hydrogen bond, called wobbling-in-the-cone motion. This motion gives rise to a fast rotational relaxation less than 1 ps.43 The longer relaxation motion represents breaking and re-forming of hydrogen bonds with other water molecules, called a jumping and frame tumbling. Rotational relaxation by jumping and frame tumbling has a different relaxation time depending on the condition. Figure 5 shows that all water molecules including bulk water molecules have fast relaxation, which results in initial decrease of reorientational time correlation function. For a longer relaxation, all rotational relaxations of buried water molecules are found to be much slower than that of bulk. Because intensity is different for each other, the availability of switching new hydrogen bond differs.44 

Performing MD simulations of WT-PrPc and T188R, we found the following two key points. (i) T188R mutation affects secondary structure leading to the increase of β-sheet content and the decrease of helix 2. (ii) Although the mutation point is located far from the buried hydration sites, the sites become unstable by the mutation. Previous studies have shown that T188 is an interface between two subdomains.45,46 We divide the prion protein into two subdomains: S1, H1 and S2 (from residue 125 to residue 171), and H2 and H3 (from residue 172 to 228). We calculate the mean distance between the centers of mass of subdomains, and the coefficient of variance (CV) [see table III]. In T188R 1, the CV is higher than those in other simulations. Because all buried water molecules connect two subdomains, the stability of the distance between two subdomains is critical for survival of the buried hydration sites. We observe a clear correlation between the CV and the mean residence time (see Fig. S1).47 It suggests that this large fluctuations of the distance make the buried hydration sites fragile.

TABLE III.

Mean distance and coefficient of variance between subdomains S1-H1-S2 and H2, 3.

Mean distance [nm] CV (10−4)
WT-PrPc 1  0.399  26.8 
WT-PrPc 2  0.429  44.2 
T188R 1  0.347  59.2 
T188R 2  0.461  26.9 
Mean distance [nm] CV (10−4)
WT-PrPc 1  0.399  26.8 
WT-PrPc 2  0.429  44.2 
T188R 1  0.347  59.2 
T188R 2  0.461  26.9 

We also focus on a configuration around the mutation. In WT-PrPc, side chain of T188 was embedded in the protein including G200. In T188R, we replaced the residue 188 from threonine to arginine. Because the side chain of arginine is larger than that of threonine, the side chain of R188 cannot be embedded in protein and exposed to the bulk in T188R 1 [see Fig. 6(A)]. However, R188 in T188R 2 was mainly connected to the side chain of G200, which may contribute to the stability of T188R 2 [see Fig. 6(B)]. In the former case, distance between the closest atoms of R188 and G200 is around 0.3 nm while that in latter case is around 0.2 nm [see Fig. 6] Figure 6 shows the distance between the closest atoms of the two residues. R188 is sometimes bent and thus the interaction between the two residues are weak. This configuration is dominant in T188R 1, while it is not in T188R 2. Moreover, the statistical test implies that the mean distance in T188R1 is significantly larger than that in T188R 2. Overall, these data support following three things: 1) T188R which is not adjacent to the buried hydration sites sometimes makes the sites fragile, 2) subdomains fluctuations are tightly correlated with the stability of the buried hydration sites, and 3) because of the size of side chain, R188 cannot be embedded in protein, and this instability may trigger the subdomains fluctuations. Our findings provide a possible explanation for the instability by mutation which is not directly adjacent to the buried hydration sites.

FIG. 6.

Distance between the closest atoms of R188 and G200. The mean distances for T188R1 and T188R2 are 0.288 nm with the standard deviation 0.04 nm and 0.249 nm with the standard deviation 0.05 nm, respectively. (A) and (B) are snapshots of T188R1 at t=450 ns and T188R2 at t=500 ns, respectively.

FIG. 6.

Distance between the closest atoms of R188 and G200. The mean distances for T188R1 and T188R2 are 0.288 nm with the standard deviation 0.04 nm and 0.249 nm with the standard deviation 0.05 nm, respectively. (A) and (B) are snapshots of T188R1 at t=450 ns and T188R2 at t=500 ns, respectively.

Close modal

In this study, we have performed all-atom MD simulations of WT-PrPc and a pathogenic point mutant T188R to investigate conformational changes and the buried hydration sites. By mutation T188R, a part of helix 2 (residue 174 to 178) is disrupted and β-sheet 2 increases in both T188R. In WT-PrPc, we have identified the four buried hydration sites, and we have found that the rotational relaxations of all buried water molecules are much slower than that of bulk water. This means that all buried water molecules are fixed in stable position by hydrogen bonds with amino acids stably. In one of T188R simulations, on the other hand, there are no stable buried hydration sites. These results show that T188R sometimes makes the buried hydration sites fragile. Subdomain analysis have revealed that the instability of the buried hydration sites is related to the subdomains fluctuations. While there is the connection between R188 and G200 in T188R 2, R188 is mainly not connected to G200 in T188R 1. These data suggest that R188 cannot be embedded due to the size of side chain and this instability may trigger subdomains fluctuations. Our results provide a possible explanation for the instability by mutation T188R.

This work was supported (in part) by MEXT Grant-in-Aid for the “Program for Leading Graduate Schools” and Keio University Program for the Advancement of Next Generation Research Projects.

2.
M.
Bucciantini
,
E.
Giannoni
,
F.
Chiti
,
F.
Baroni
,
L.
Formigli
,
J.
Zurdo
,
N.
Taddei
,
G.
Ramponi
,
C.
Dobson
, and
M.
Stefani
,
Nature
416
,
507
(
2002
).
3.
M. I.
Apostol
,
K.
Perry
, and
W. K.
Surewicz
,
J. Am. Chem. Soc.
10202
(
2013
).
4.
R.
Zahn
,
A.
Liu
,
T.
Lührs
,
R.
Riek
,
C.
Von Schroetter
,
F.
López García
,
M.
Billeter
,
L.
Calzolai
,
G.
Wider
, and
K.
Wüthrich
,
Proc. Nat. Acad. Sci. U.S.A.
97
,
145
(
2000
).
5.
K.
Pan
,
M.
Baldwin
,
J.
Nguyen
,
M.
Gasset
,
A.
Serban
,
D.
Groth
,
I.
Mehlhorn
,
Z.
Huang
,
R.
Fletterick
, and
F.
Cohen
,
Proc. Nat. Acad. Sci. U.S.A.
90
,
10962
(
1993
).
6.
W.
Swietnicki
,
R.
Petersen
,
P.
Gambetti
, and
W.
Surewicz
,
J. Biol. Chem.
272
,
27517
(
1997
).
7.
S.
Hornemann
and
R.
Glockshuber
,
Proc. Nat. Acad. Sci. U.S.A
95
,
6010
(
1998
).
8.
Y.
Matsunaga
,
D.
Peretz
,
A.
Williamson
,
D.
Burton
,
I.
Mehlhorn
,
D.
Groth
,
F. E.
Cohen
,
S. B.
Prusiner
, and
M. A.
Baldwin
,
Proteins
44
,
110
(
2001
).
9.
R.
Gerber
,
A.
Tahiri-Alaoui
,
P.
Hore
, and
W.
James
,
Protein Sci.
17
,
537
(
2008
).
10.
E.
El-Bastawissy
,
M.
Knaggs
, and
I.
Gilbert
,
J. Mol. Graph. Model.
20
,
145
(
2001
).
11.
R.
Riek
,
G.
Wider
,
M.
Billeter
,
S.
Hornemann
,
R.
Glockshuber
, and
K.
Wüthrich
,
Proc. Nat. Acad. Sci.
95
,
11667
(
1998
).
12.
M.
Van Der Kamp
and
V.
Daggett
,
Protein. Eng. Des. Sel.
22
,
461
(
2009
).
13.
X.
Chen
,
S.
Zhu
,
S.
Wang
,
D.
Yang
, and
J.
Zhang
,
Molec. Simul.
1
(
2013
).
14.
A.
Apetri
,
K.
Surewicz
, and
W.
Surewicz
,
J Biol. Chem.
279
,
18008
(
2004
).
15.
N.
Chakroun
,
A.
Fornili
,
S.
Prigent
,
J.
Kleinjung
,
C. A.
Dreiss
,
H.
Rezaei
, and
F.
Fraternali
,
J. Chem. Theory Comput.
9
,
2455
(
2013
).
16.
C.
Schröder
,
T.
Rudas
,
S.
Boresch
, and
O.
Steinhauser
,
The Journal of chemical physics
124
,
234907
(
2006
).
17.
T.
Angel
,
M.
Chance
, and
K.
Palczewski
,
Proc. Nat. Acad. Sci.
106
,
8555
(
2009
).
18.
N.
Takano
,
K.
Umezawa
,
J.
Ikebe
,
Y.
Sonobe
,
R.
Yagisawa
,
J.
Ito
,
N.
Hamasaki
,
D.
Mitomo
,
H.
Miyagawa
,
A.
Yamagishi
 et al,
Chem. Biol. Info. J.
8
,
14
(
2008
).
19.
S.
Chong
and
S.
Ham
,
Proc. Nat. Acad. Sci. U.S.A
109
,
7636
(
2012
).
20.
M.
Jana
and
S.
Bandyopadhyay
,
The Journal of chemical physics
137
,
055102
(
2012
).
21.
Y.
Lu
,
C.-Y.
Yang
, and
S.
Wang
,
J. Am. Chem. Soc.
128
,
11830
(
2006
).
22.
E.
Yamamoto
,
T.
Akimoto
,
Y.
Hirano
,
M.
Yasui
, and
K.
Yasuoka
,
Physical Review E
87
,
052715
(
2013
).
23.
E.
Yamamoto
,
T.
Akimoto
,
M.
Yasui
, and
K.
Yasuoka
,
Scientific reports
4
(
2014
).
24.
R.
Malin
,
P.
Zielenkiewicz
, and
W.
Saenger
,
J. Biol. Chem.
266
,
4848
(
1991
).
25.
U.
Sreenivasan
and
P. H.
Axelsen
,
Biochemistry
31
,
12785
(
1992
).
26.
S.
Shaltiel
,
S.
Cox
, and
S. S.
Taylor
,
Proc. Nat. Acad. Sci. U.S.A
95
,
484
(
1998
).
27.
J.
Knight
,
D.
Hamelberg
,
J.
McCammon
, and
R.
Kothary
,
Proteins
76
,
527
(
2009
).
28.
L.
Haire
,
S.
Whyte
,
N.
Vasisht
,
A.
Gill
,
C.
Verma
,
E.
Dodson
,
G.
Dodson
, and
P.
Bayley
,
J. Mol. Biol.
336
,
1175
(
2004
).
29.
A.
De Simone
,
G.
Dodson
,
C.
Verma
,
A.
Zagari
, and
F.
Fraternali
,
Proc. Nat. Acad. Sci. U.S.A
102
,
7535
(
2005
).
30.
H.
Lorenz
,
O.
Windl
, and
H. A.
Kretzschmar
,
J. Biol. Chem.
277
,
8508
(
2002
).
31.
S.
Roeber
,
E.-M.
Grasbon-Frodl
,
O.
Windl
,
B.
Krebs
,
W.
Xiang
,
C.
Vollmert
,
T.
Illig
,
A.
Schröter
,
T.
Arzberger
,
P.
Weber
 et al,
PLoS One
3
,
e2147
(
2008
).
32.
J.
Guo
,
L.
Ning
,
H.
Ren
,
H.
Liu
, and
X.
Yao
,
Biochim. Biophys. Acta.
(
2011
).
33.
L.
Calzolai
and
R.
Zahn
,
J. Biol. Chem.
278
,
35592
(
2003
).
34.
N.
Guex
and
M. C.
Peitsch
,
Electrophoresis
18
,
2714
(
1997
).
35.
J.
Phillips
,
R.
Braun
,
W.
Wang
,
J.
Gumbart
,
E.
Tajkhorshid
,
E.
Villa
,
C.
Chipot
,
R.
Skeel
,
L.
Kale
, and
K.
Schulten
,
J. Comput. Chem.
26
,
1781
(
2005
).
36.
V.
Hornak
,
R.
Abel
,
A.
Okur
,
B.
Strockbine
,
A.
Roitberg
, and
C.
Simmerling
,
Proteins
65
,
712
(
2006
).
37.
W.
Kabsch
and
C.
Sander
,
Biopolymers
22
,
2577
(
1983
).
38.
R.
Dima
and
D.
Thirumalai
,
Biophysical journal
83
,
1268
(
2002
).
39.
R. I.
Dima
and
D.
Thirumalai
,
Proceedings of the National Academy of Sciences of the United States of America
101
,
15335
(
2004
).
40.
X.
Lu
,
P. L.
Wintrode
, and
W. K.
Surewicz
,
Proceedings of the National Academy of Sciences
104
,
1510
(
2007
).
41.
A.
De Simone
,
G.
Dodson
,
F.
Fraternali
, and
A.
Zagari
,
FEBS Lett.
580
,
2488
(
2006
).
42.
D.
Laage
and
J.
Hynes
,
Science
311
,
832
(
2006
).
43.
D.
Laage
and
J.
Hynes
,
J. Phys. Chem. B
112
,
14230
(
2008
).
44.
D.
Laage
and
J.
Hynes
,
Chem. Phys. Lett.
433
,
80
(
2006
).
45.
A.
De Simone
,
A.
Zagari
, and
P.
Derreumaux
,
Biophys. J.
93
,
1284
(
2007
).
46.
S.
Schwarzinger
,
A. H.
Horn
,
J.
Ziegler
, and
H.
Sticht
,
J. Biomol. Struct. Dyn.
23
,
581
(
2006
).
47.
See supplementary material at http://dx.doi.org/10.1063/1.4953061 for the figure of normalized CV and normalized residence time.

Supplementary Material