ARTICLE pubs.acs.org/JPCB
Methanol Strengthens Hydrogen Bonds and Weakens Hydrophobic Interactions in Proteins A Combined Molecular Dynamics and NMR study Soyoun Hwang,†,§ Qiang Shao,‡,§ Howard Williams,† Christian Hilty,†,* and Yi Qin Gao‡,* † ‡
Center for Biological NMR, Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States College of Chemistry and Molecular Engineering, Beijing National Laboratory of Molecular Sciences, Peking University, Beijing 100871, China ABSTRACT: A combined simulation and experimental study was performed to investigate how methanol affects the structure of a model peptide BBA5. BBA5 forms a stable β-hairpin-R-helix structure in aqueous solutions. Molecular dynamics simulations were performed in water and methanol/water solutions using allatom explicit models. NMR experiments were used to test the calculated results. The combined theoretical and experimental studies suggest that methanol strengthens the interactions between the polar backbone of the peptide and thus enhances the secondary structure formation; at the same time methanol weakens the hydrophobic interactions and results in an expansion of the hydrophobic core and an increase in gyration.
’ INTRODUCTION Protein structure can be investigated with increasing ease and precision using the modern techniques of structural biology. Not surprisingly, research focus has shifted to the study of the structure and function of large proteins, protein complexes, or even membrane embedded proteins; yet, even the basic determinants of protein structure and folding are still poorly understood. In addition to intramolecular interactions, protein structure appears to be governed by contributions from solvation effects. The importance of the latter is illustrated by the effects of the interaction with various solutes, leading to the protein folding or denaturation. Observing interactions between solvent species and a protein remains challenging using the currently available experimental techniques for various reasons. For example, X-ray crystallography requires the artificial environment of a crystal, and furthermore cannot observe disordered conformations, whereas in nuclear magnetic resonance (NMR) spectroscopy dynamic effects reduce the ability to observe intermolecular solvent/ protein interactions. However, the techniques of computational chemistry give access to a complete set of molecular parameters, in addition to the high time resolution, and they present a unique opportunity to further the understanding of the processes governing protein structure and function. The interpretation of computational results without physical data is often problematic. However, if results are compared to experimentally determined structures, they become capable of filling in mechanistic information that experimental structures alone cannot provide. The solvation effects during protein folding remain one of the most intriguing problems in physical chemistry. Examination of protein structural changes occurring as a result of the changes in r 2011 American Chemical Society
solvation environment can serve as a useful model system for understanding the various molecular interactions in determining protein structures. A typical case is the alcohol-induced conformational change of protein: it was observed in various experiments that the addition of alcohol into the aqueous solution generates the structure of protein similar to the molten globule (H) state, a common folding intermediate state for small globular proteins.1,2 As a result, alcohols (particularly methanol and trifluoroethanol (TFE)) have been widely used in the protein folding and structure investigations, the experimental approaches used include NMR, circular dichroism (CD), Fourier transform infrared spectroscopy (FTIR), light scattering, fluorescence, and so on.215 The previous experimental and theoretical studies have shown that the addition of methanol to aqueous solutions of proteins stabilizes (or even induces) the R-helical structure and can also at the same time denature other protein structures, accompanied by the accumulation of methanol near the protein surfaces.2,14,16 Recent molecular dynamics (MD) simulations combined with NMR studies are largely consistent with these studies.17 In this article, we report a combined MD simulation and NMR experimental study to show the details how methanol affects the structure and various molecular interactions in BBA5. The BBA5 peptide is from a family of peptides designed to exhibit a ββR fold in water: BBA5 forms a stable tertiary structure consisting of a β-hairpin segment (Tyr1—Phe8) and an R-helix segment (Arg10—Gly23); its globular structure is stabilized by the Received: December 2, 2010 Revised: March 14, 2011 Published: May 02, 2011 6653
dx.doi.org/10.1021/jp111448a | J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B hydrophobic core cluster constructed by Tyr1, Val3, Tyr6, and Phe8 from the hairpin segment and Leu14, Leu17, and Leu18 from the helix segment.18 Because of its well-defined secondary structures, BBA5 represents an ideal model to test the stability of secondary structures.19 In this study, the computation was first used to make predictions, and NMR and CD experiments were then used to test the calculated results. These studies show that the change of the solvation environment has a strong effect on the structure of BBA5: the backbone hydrogen bond interaction is strengthened, and the side-chain hydrophobic interaction is weakened in the presence of methanol. As a result, the solution structure has significantly more helical components, whereas the overall structure becomes more open.
’ EXPERIMENTAL METHODS Sample Preparation. BBA5 (NH2YRVDPSYDFSRSDE
LAKLLRQHAG-COO) peptide samples were obtained from solid-phase synthesis (Anaspec, Fremont, CA) and used without further purification.18 The lyophilized powder was dissolved in 90% H2O/10% D2O, or in 50% (v/v) deuterated methanol/ water (90% H2O/10% D2O), to a peptide concentration of 1 mM. Buffer was not used and the pH was adjusted to pH 4.5. For CD spectroscopy, samples were further diluted to a concentration of 20 μM. Concentrations were determined by spectrophotometry, using a calculated molar extinction coefficient ε = 2980 M1cm1. Structure Determination. CD spectra were acquired using an AVIV 26DS spectrometer, from 190 to 260 nm in steps of 1 nm, at a temperature of 280 K. R-Helix content was estimated from mean residue ellipticity at 222 nm,20 neglecting any contributions from β-sheet-like secondary structure to the CD signal. Solution NMR structures of the peptides were determined using 2D 1H NMR.21 NMR spectra were measured on a Varian INOVA 600 MHz spectrometer with triple resonance probe and z-gradient, and on a Bruker ARX 500 MHz spectrometer with triple resonance probe and z-gradient. All spectra were measured at 280 K. Amino acid spin systems were identified in TOCSY spectra (70 ms mixing time),22 and sequence specific resonance assignments were obtained from NOESY spectra (150 ms mixing time).23 All of the NOE assignments and structure calculations were performed by the same automated procedure using the program CYANA,24 to ensure objective results. In each of seven cycles, distance constraints were automatically identified in the NOESY spectra, and were used for a structure calculation by simulated annealing starting from 100 randomized structures. From the last cycle, a bundle of the 10 conformers with lowest target function were retained to represent the NMR structure. Simulation Details. MD simulations on BBA5 in pure water and methanol/water (MeOH/water) solution were performed in explicit solvent making use of AMBER 9.0 suite of programs25 with FF99 force field.26 Water is described with SPC/E model27 and the parameters of the force field concerning the methanol are taken from Caldwell et al.28 The initial structure of BBA5 was taken from the NMR structure (PDB code: 1T8J18). The system of BBA5 in MeOH/water solution was prepared by immersing BBA5 peptide into a cubic box containing 800 methanol and 2509 water molecules. Meanwhile, the system of BBA5 in pure water was prepared by immersing BBA5 peptide into a cubic box containing 3413 water molecules. In both systems, one Cl- anion was added to neutralize the charge. Cubic boxes were created
ARTICLE
using LEaP program in AMBER 9.0, with water and other solvent molecules evenly distributed in the initial configuration. For each system, the simulation procedure includes the energy minimization, the following heating-up process, and the final longtime equilibrium simulation calculation (production). NPT (constant number, pressure, temperature) ensemble calculations were performed and the periodic boundary conditions were used in the simulations. The SHAKE algorithm29 was used to constrain all bonds involving hydrogens. A cutoff of 10.0 Å was applied for nonbonding interactions. The Particle Mesh Ewald method was applied to treat long-range electrostatic interactions.30 Initially, the energies of the systems were minimized through a total of 2500 steps of calculations: the first 1000 steps are the steepest descent minimization with the peptide being fixed using harmonic restraints, with a force constant of 500.0 kcal mol1 Å2 applied to the backbone atoms, the following 1500 steps are the conjugate gradient minimization. Subsequently, the systems were heated to 360 K within 200 ps and equilibrated at 360 K for 1 ns, followed by a 200 ps of cooling from 360 to 300 K (these heating, equilibration and cooling processes were run with harmonic restraints applied to the backbone atoms (force constant =10.0 kcal mol1 Å2); the equilibration at high temperature makes the peptide and methanol molecules dissolved in water). Finally, the production runs of 224 ns for BBA5 in MeOH/water solution and of 230 ns for BBA5 in pure water were performed and then used for the data analysis, with the data being collected every 1.0 ps. Secondary structures were analyzed by STRIDE.31 Hydrogen Bond Definition in MD Simulations. A hydrogen bond is considered as formed only if the distance between the hydrogen donor and acceptor is less than 3.5 Å and the NHO (or OHO) angle is greater than 135°.
’ RESULTS AND DISCUSSION CD spectroscopy of BBA5 samples indicated that the peptide adopts a predominantly R-helical structure both in water and in 50% (v/v) MeOH/water solutions, under the experimental conditions used (Figure 1). As judged from the ellipticity at 222 nm, the R-helix content in MeOH/water solution is significantly increased (Table 1). More detailed structure information is available from the NMR experiments. NOE distance constraints were collected from NOESY spectra (Figure 2). On the basis of these constraints, structures of BBA5 in water and in
Figure 1. CD spectra of BBA5 in water (closed circle), and MeOH/ water solution (open circle) at a measurement temperature 280 K. 6654
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B
ARTICLE
Table 1. Mean Residue Ellipticity at 222 nm (θ222) and Estimated r-Helix Content from CD Spectra of BBA5a In MeOH/Water
a
In Water
T [K]
θ222 [ 103 deg 3 cm2 3 dmol1]
estimated R-helix content
θ222 [ 103 deg 3 cm2 3 dmol1]
280
20 582.92
56%
10 662.21
30%
288
19 928.82
55%
9612.04
27%
298
17 878.03
49%
9297.66
26%
308
17 039.25
47%
8229.65
24%
318
15 184.69
42%
7560.76
22%
328
13 487.88
38%
7157.19
21%
estimated R-helix content
Estimates of R-helix content are based on θ222 and neglect any contribution of β-sheet like secondary structure to the CD signal.
Figure 2. BBA5 in a) water, and b) MeOH/water solution. The top panels indicate observed NOEs in the respective spectra. The bottom panels show the deviation of HR chemical shift from random coil values. SC denotes side-chain.
Figure 3. NMR structure of BBA5 in a) water, and b) MeOH/water solution. The side-chains of hydrophobic residues participating in the hydrophobic cluster in water are indicated.
MeOH/water solutions were calculated (Figure 3). The NMR structure in water is similar to a previously determined structure, with a few small differences mainly in the side-chain packing.18 In the following comparisons, we use the newly calculated structure in order to avoid the bias that may otherwise arise due to the differences in methodology. In water, dRβ(i,iþ3) and dRN(i,iþ3) NOEs, indicative of helical secondary structure, were observed between residues R10 and A22
Figure 4. Regions taken from NOESY spectra of BBA5 in a) water, and b) MeOH/water solution. NOE crosspeaks corresponding to longrange interactions are shown.
(Figure 3). The helix extent in MeOH/water solution was slightly larger (S9 to G23), and overall more medium-range NOEs were observed, indicating an increased propensity for the R-helix. This observation is in agreement with the results from CD spectroscopy. With other peptides, alcohols have also been observed to increase R-helix content, an effect presumed to arise because of a reduced number of hydrogen bonds between the solvent and the peptide backbone when compared to water.14,3234 6655
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B
ARTICLE
The β-hairpin-like element in water is defined by cross-strand NOEs between V3 and F8. Additionally, a number of long-range interactions with the R-helix indicate a tertiary contact (Y1L14, Y6L14, and F8L14). These NOEs are part of a hydrophobic cluster, which appears to stabilize the β-hairpin structure. In Table 2. Average Number of Individual Backbone Hydrogen Bonds of BBA5 Formed in Water, and MeOH/Water Solutions Respectively from MD Simulation Dataa
a
in MeOH/
in
backbone
in MeOH/
in
backbone HB
water
water
HB
water
water
F8OY1H
0.000
0.000
E13OL17H
0.326
0.000
Y1OF8H
0.590
0.603
L14OL18H
0.335
0.492
Y6OV3H
0.950
0.978
A15OR19H
0.252
0.222
V3OY6H R10OL14H
0.204 0.311
0.282 0.000
K16OQ20H L17OH21H
0.325 0.044
0.029 0.265
S11OA15H
0.416
0.188
L18OA22H
0.270
0.541
D12OK16H
0.363
0.000
R19OG23H
0.273
0.003
The average is over the entire simulation trajectory.
Figure 5. Time evolution of the number of native backbone hydrogen bonds in a) β-hairpin, b) R-helix segments of BBA5, and (c) the total number of backbone hydrogen bonds from MD simulation data (black, BBA5 in MeOH/water solution; red, BBA5 in water).
contrast to its effect on the R-helix, MeOH/water solution appears to destabilize the β-hairpin-like structure. Fewer cross strand NOEs are observed, and the tertiary interaction is lost (Figure 2). A direct, visual indication of the loss of these contacts is also obtained by comparing the NOESY spectra acquired under the two conditions, shown in Figure 4. A significant reduction in the number of crosspeaks involving the side-chains of aromatic residues in the hydrophobic core cluster can be seen in MeOH/water solution. To compare MD simulations with experiments, we show in Table 2 the population of the individual backbone hydrogen bond (HB)s calculated from the simulations. Additionally, the time evolution of the number of hairpin and helical hydrogen bonds is shown in Figure 5 and the time evolution of the secondary structures is shown in Figure 6. It is apparent that the addition of methanol increases the population of the protein backbone hydrogen bonds, in particular those belonging to the R-helix. This effect of methanol is more evident in the last stage of the simulations as shown in the two figures. The average number of R-helix hydrogen bonds in MeOH/water solution is about 1.7 times of that in pure water. This number is similar to the ratio of R-helical secondary structure content estimated from the CD spectra in MeOH/water and water samples, which is about 1.9 at the room temperature (Table 1). Although the hydrogen bond numbers calculated from the simulations does not compare directly to the helicity in CD experiment in a quantitative manner, these results show that the helix formation is enhanced in MeOH/water solution as compared to in pure water. In addition to hydrogen bonding information, we also evaluated the overall structure of BBA5 obtained from MD simulations. Parts a and b of Figure 7 show the time evolution of the solvent accessible surface area (SASA), and the radius of gyration (Rg) of the overall BBA5 in water and MeOH/water solutions obtained from MD simulations respectively. Smaller SASA and Rg indicate a more compact peptide structure in water. It is also seen from this figure that the corresponding values increase in the presence of methanol. These observations indicate that methanol enhances the exposure of the protein to the solvent. To
Figure 6. Secondary structure as a function of MD simulation time for BBA5 in a) MeOH/water solution and b) water. β-strand, R-helix, 310-helix, turn, and coil are shown in yellow, purple, blue, cyan, and white, respectively. 6656
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B understand which part of BBA5 contributes most significantly to this structural exposure in MeOH/water solution, we calculated the radius of gyration of the hydrophobic core cluster formed between the β-hairpin and R-helix segments (Rgcore), and that of the two segments respectively (Rg, hairpin and Rg, helix). As shown in part c of Figure 7, Rg, hairpin and Rg, helix remain steady during the simulation, whereas Rgcore increases in a synchronous manner with Rg. Therefore, it is the breaking of the hydrophobic core cluster between the β-hairpin and R-helix segments that is responsible for the expansion of BBA5. This result corresponds to the observed loss of the tertiary contact between the β-hairpin and R-helix segments in MeOH/water solution in the NMR experiment, as discussed above. To further assess the propensity of the individual segments of the peptide to form regular secondary structures, the temperature
Figure 7. Time evolution of a) the solvent accessible surface area (SASA), b) the radius of gyration (Rg) of overall BBA5 peptide, and c) the radius of gyration of the hydrophobic core packed between the R-helix and β-hairpin segments (Rgcore) of BBA5 in MeOH/water solution (black) and pure water (red) from MD simulation data.
ARTICLE
coefficients for amide proton chemical shifts (ΔσHN/ΔT) were determined for each sample (parts a and b of Figure 8). Values of ΔσHN/ΔT < 4.6 ppb/K are generally taken as an indication of the presence of an intramolecular hydrogen bond, where the amide proton is protected from the solvent.35 For BBA5 in water, the C-terminal part exhibits a significant number of values smaller than 4.6 ppb/K, indicating the presence of the R-helix. The N-terminal segment, where the β-hairpin structure is located, shows the temperature coefficient values predominantly larger than 4.6 ppb/K. This difference may be rationalized considering that for β-hairpin structure in BBA5, the stabilization by the hydrophobic core cluster may play a more important role than the hydrogen bonding. In MeOH/water solution, the R-helix region (A15 Q20) exhibits unusually large temperature coefficients. Such an effect has previously been observed for peptides solvated in TFE, where it was hypothesized that it could be due to the hydrophobic interactions between the helix and the organic solvent.36 It may be of interest to further examine this point based on the simulation results. In Figure 9, we show the effects of methanol on the hydrogen bonds between protein backbone and solvent molecules, for each residue in the simulations (in MeOH/water solution, the hydrogen bonds between the backbone and solvent include not only those from water molecules but also those from methanol molecules, both are calculated and shown in Figure 9). For most of the residues, the hydrogen bonding from water is largely reduced by the addition of methanol, which is to some extent compensated by the hydrogen bonding from methanol. Meanwhile, the overall hydrogen bonding between protein backbone and solvent including water and methanol is only slightly changed compared to that in pure water (the total number of hydrogen bonds between the backbone of the 23 residues of BBA5 and solvent averaged over the simulation trajectory is 23.53 in MeOH/water and 25.45 in water). Part d of Figure 8 shows the difference in the average number of the hydrogen bonds from solvent for each residue between
Figure 8. a) Amide proton temperature coefficients (ΔσHN/ΔT) for BBA5 in water. b) Temperature coefficients for BBA5 in MeOH/water solution. The dashed line at ΔσHN/ΔT = 4.6 ppb/K indicates a typical cutoff value for the identification of secondary structure elements (see text). c) Difference in temperature coefficients for BBA5 between water and MeOH/water solutions, Δ(ΔσHN/ΔT) = (ΔσHN, water/ΔT) (ΔσHN, methanol/ΔT). d) Difference in the average number of hydrogen bonds from solvent molecules, between MD simulations of BBA5 in water and in MeOH/water solutions, Δn = (HBs from water)in water (HBs from water þ HBs from methanol)in MeOH/water. 6657
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B
ARTICLE
simulations for BBA5 in water and MeOH/water solutions. The hydrogen bonding change between the two simulation systems varies with the residue number. Most noticeably, the number of hydrogen bonds decreases significantly around residue R10 and R19 and increases for residue L17 in MeOH/water solution. This result correlates with the observed change in temperature coefficients from the NMR measurement (part c of Figure 8). In MeOH/water solution, the small change of the total number of the proteinsolvent hydrogen bonds but the significant decrease of the total number of proteinwater hydrogen bonds, accompanying with the structural exposure of BBA5, might be attributed to the methanol accumulation near the protein surface and the consequent replacement of the water molecules. The preferred binding of methanol to the protein is calculated as the ratio RMW between the numbers of methanol and water molecules within 5 Å of the protein surface and then normalized according to the total numbers of methanol (800) and water (2509) molecules in the simulation solution. The resulting values of the normalized MeOH/water ratios are greater than 1 for all residues as shown in Table 3, and the
average of RMW over all residues is 1.39, indicating the preferred binding of methanol over water to the protein surface. The more favored interaction between methanol and the protein reduces the hydrophobic effect and therefore results in a higher degree of solvent exposure of the peptide, as indicated by the increase of SASA (Figure 7). Meanwhile, methanol is a weaker hydrogen bond forming agent than water: as shown in Figure 10, a significantly higher peak is present at the characteristic hydrogen bond distance for OO or NO (2.5 to 3 Å) in the water peptide radial distribution functions than in the corresponding one for methanol. Although the structure of BBA5 is expanded in MeOH/water solution and the contact area for solvent is thus increased, the accumulation of methanol rather than water near the protein surface cannot largely increase the proteinsolvent hydrogen bonding. However, the accumulation of methanol in general reduces the proteinwater hydrogen bonding as well as the local dielectric constant near protein surface. This change is not uniform
Figure 9. Average number of hydrogen bonds formed between the backbone amide (NH) or carbonyl (CdO) group of individual residues with solvent molecules from MD simulation data. The numbers of hydrogen bond formed are shown in black for those between individual residues and water, in red for those between individual residues and methanol in MeOH/water solution, and in green for those between individual residues and water in pure water. The average is over the entire simulation trajectory.
Figure 10. Radial distribution functions (g(r)) for different groups of a) methanol, and b) water molecules around peptide backbone CO (black) and NH (red) groups at 716 ns (solid line), 195204 ns (dashed line) of the trajectory in MD simulation of BBA5 in MeOH/water solution. OBOM, methanol oxygen atoms around peptide backbone CO groups; NBOM:, methanol oxygen atoms around peptide backbone NH groups; OBOW:, water oxygen atoms around peptide backbone CO groups; NBOW, water oxygen atoms around peptide backbone NH groups.
Table 3. Average Number of Solvent Molecules within 5 Å of Each Individual Residue of BBA5, from MD Simulation Dataa Residue Number
In MeOH/Water
In Water
water
MeOH
RMW
Residue Number
In MeOH/Water
In Water
water
MeOH
RMW
Y1
66.35
25.54
1.21
67.02
E13
42.47
15.73
1.16
R2
58.71
22.01
1.18
54.02
L14
15.94
12.52
2.46
50.68 22.16
V3
17.08
10.80
1.98
16.88
A15
14.48
7.89
1.71
27.26
P4
27.57
12.97
1.48
36.36
K16
43.51
15.57
1.12
54.06
S5
29.28
13.34
1.43
34.51
L17
22.86
13.37
1.83
29.40
Y6 D7
34.96 27.70
21.24 13.79
1.90 1.56
32.11 22.90
L18 R19
21.75 50.35
14.08 19.94
2.03 1.24
22.20 66.12
F8
35.82
20.45
1.79
39.36
Q20
41.18
18.06
1.38
46.88
S9
30.35
12.46
1.29
35.76
H21
44.76
19.88
1.39
42.45
R10
55.20
24.70
1.40
65.96
A22
22.05
10.43
1.48
28.34
S11
18.43
13.08
2.22
24.82
G23
36.48
12.45
1.07
39.37
D12
44.61
15.24
1.07
48.13
The first two data columns are for BBA5 in MeOH/water solution, and the fourth column is for BBA5 in water. The average is over the entire simulation trajectory. RMW (the third column) is the ratio between the numbers of methanol and water molecules accumulated around each residue, which is then normalized according to the total numbers of methanol and water molecules in the simulation solution. a
6658
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B along the protein: the reduction of proteinwater interaction and the increase of the proteinprotein hydrogen bonding mainly occur in the R-helix region, as shown in Figure 9 and Table 2, respectively. The overall decrease of proteinwater hydrogen bonding for residues from R10 to R19 correlates well with the increase of the corresponding proteinprotein hydrogen bonding, revealing the fact that the source of the proteinprotein hydrogen bond disruption is from the water hydrogen bonding. Furthermore, even along the helix portion, the change in proteinwater hydrogen bonding also varies with the sequence. In fact, the proteinwater hydrogen bonding is mostly weakened for the polar residues (e.g., R10, S11, D12, K16, and R19) but little changed or even strengthened for the apolar ones (e.g., L14, A15, L17, and L18). Consequently, the proteinprotein hydrogen bonds formed by the carbonyl groups from polar residues (R10OL14H, S11OA15H, D12OK16H, E13OL17H, K16OQ20H, and R19OG23H) are strengthened, whereas the stability of the hydrogen bonds formed by the carbonyl groups from apolar residues (L14OL18H, A15OR19H, L17OH21H, and L18OA22H) are either weakly affected or decreased. These results are consistent with the difference between the temperature coefficients obtained in pure water and in MeOH/water solution: Figure 8 indicates that the helix becomes more structured in the position where the polar residues are gathered, for example, the N-terminal segment (residues R10E13), rather than the more apolar part (A15R19) in MeOH/water solution. It is also interesting to point out that the residues that form the better secondary structure are more likely to hydrogen bond with methanol than with water when compared to those associated with broken secondary structures, as shown by the ratio between the numbers of their hydrogen bonds to methanol and water (Figure 9). This observation indicates that proteinwater hydrogen bonds are likely more capable of breaking the protein secondary structure than those formed between protein and methanol.
ARTICLE
although the introduction of methanol increases the exposure of the protein to the solvent. The effects of methanol on the secondary structure are not uniform along the R-helix but dependent on the sequence and/or the residue position: the backbone hydrogen bonds formed by the carbonyl groups from polar residues are strengthened whereas those from apolar residues are weakened. The effect of methanol on the protein structure is thus a combination of direct (preferred binding of methanol) and indirect (e.g., reduced proteinwater hydrogen bonding) effects, although these two are closely related, and manifested differently in local and global structures. The balance between these effects determines whether methanol tightens or loosens the local protein structure, which heavily depends on the local sequence and environment.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected] (C.H.);
[email protected] (Y.Q.G.). §
Both authors contributed equally to this work.
’ ACKNOWLEDGMENT C.H. thanks the Camille and Henry Dreyfus Foundation for a New Faculty Award. Support from the Welch Foundation (Grant A-1658), from the National Science Foundation (Grant CHE0846402), and from Texas A&M University startup funds is gratefully acknowledged. NMR measurements at 600 MHz were performed in the Texas A&M Biomolecular NMR Laboratory, and measurements at 500 MHz in the Texas A&M Chemistry NMR facility. We thank Dr. Xiangming Kong for assistance with NMR measurements. Y.Q.G. is a 2006 Searle Scholar and a 2008 Changjiang Scholar. ’ REFERENCES
’ CONCLUSIONS In this article, we performed MD simulation and NMR/CD studies to investigate the effects of methanol on a short peptide that serves as a model system for proteins. The combined theoretical and experimental data show that the addition of methanol enhances the formation of secondary structure, especially the R-helical segment of the BBA5 peptide. This enhancement of backbonebackbone interactions is accompanied by the weakening of hydrophobic interactions between the side-chains, which disrupts a hydrophobic core cluster in BBA5. The peptide structure in MeOH/water solution is therefore expanded, but contains highly ordered secondary structures, consistent with earlier studies.16 The present study suggests that the accumulation of methanol near the protein surface induces the expansion of protein structure, presumably by the reduction of hydrophobic effects. At the same time, the replacement of water molecules from the local environment of the protein surface decreases the hydrogen bonding of water to protein and on average increases the proteinprotein hydrogen bonds. The hydrogen bonding of methanol to the protein backbone, on a per molecule basis, is less favorable than the hydrogen bonding of water. These effects together locally decrease the polar interaction between the solvent and protein and therefore result in a higher propensity for secondary structure formation, in particular for the R-helix,
(1) Uversky, V. N.; Narizhneva, N. V.; Kirschstein, S. O.; Winter, S.; Lober, G. Fold. Des. 1997, 2, 163–172. (2) Buck, M. Q. Rev. Biophys. 1998, 31, 297–355. (3) Harding, M. M.; Williams, D. H.; Woolfson, D. N. Biochemistry 1991, 30, 3120–3128. (4) Cox, J. P. L.; Evans, P. A.; Packman, L. C.; Williams, D. H.; Woolfson, D. N. J. Mol. Biol. 1993, 234, 483–492. (5) Albert, J. S.; Hamilton, A. D. Biochemistry 1995, 34, 984–990. (6) Babu, K. R.; Douglas, D. J. Biochemistry 2000, 39, 14702–14710. (7) Gast, K.; Siemer, A.; Zirwer, D.; Damaschun, G. Eur. Biophys. J. Biophys. 2001, 30, 273–283. (8) Kinoshita, M.; Okamoto, Y.; Hirata, F. J. Am. Chem. Soc. 2000, 122, 2773–2779. (9) Zhao, J. H.; Liu, H. L. Chem. Phys. Lett. 2006, 420, 235–240. (10) Perham, M.; Liao, J.; Wittung-Stafshede, P. Biochemistry 2006, 45, 7740–7749. (11) Kony, D. B.; Hunenberger, P. H.; van Gunsteren, W. F. Protein Sci. 2007, 16, 1101–1118. (12) Hamada, D.; Kuroda, Y.; Tanaka, T.; Goto, Y. J. Mol. Biol. 1995, 254, 737–746. (13) Yamazaki, K.; Iwura, T.; Ishikawa, R.; Ozaki, Y. J. Biochemistry 2006, 140, 49–56. (14) Hirota, N.; Mizuno, K.; Goto, Y. J. Mol. Biol. 1998, 275, 365–378. (15) Roccatano, D.; Colombo, G.; Fioroni, M.; Mark, A. E. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12179–12184. (16) Timasheff, S. N. Annu. Rev. Bioph. Biom. 1993, 22, 67–97. (17) Wang, D. Q.; Jaun, B.; van Gunsteren, W. F. Chembiochem 2009, 10, 2032–2041. 6659
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
The Journal of Physical Chemistry B
ARTICLE
(18) Struthers, M.; Ottesen, J. J.; Imperiali, B. Fold. Des. 1998, 3, 95–103. (19) Rogers, J. M. G.; Lippert, L. G.; Gai, F. Anal. Biochem. 2010, 399, 182–189. (20) Rohl, C. A.; Chakrabartty, A.; Baldwin, R. L. Protein Sci. 1996, 5, 2623–2637. (21) W€uthrich, K. NMR of Proteins Nucleic. Acids; Wiley: New York, 1986. (22) Braunschweiler, L.; Ernst, R. R. J. Magn. Reson. 1983, 53, 521–528. (23) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys. 1979, 71, 4546–4553. (24) G€untert, P.; Mumenthaler, C.; W€uthrich, K. J. Mol. Biol. 1997, 273, 283–298. (25) Case, D. A. D., T., A.; Cheatham, T. E. III; Simmerling, C. L.; Wang, j.; Duke, R. E.; Luo, R.; Merz, K. M.; Pearlman, D. A.; Crowley, M.; Walker, R. C.; Zhang, W.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra, G.; Wong, K. F.; Paesani, F.; Wu, X.; Brozell, S.; Tsui, V.; Gohlke, H.; Yang, L.; Tan, C.; Mongan, J.; Hornak, V.; Cui, G.; Beroza, P.; Mathews, D. H.; Schafmeister, C.; Ross, W. S.; Kollman, P. A. AMBER 9; University of California: San Francisco, 2006. (26) Wang, J. M.; Cieplak, P.; Kollman, P. A. J. Comput. Chem. 2000, 21, 1049–1074. (27) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. J. Phys. Chem. 1987, 91, 6269–6271. (28) Caldwell, J. W.; Kollman, P. A. J. Phys. Chem. 1995, 99, 6208–6219. (29) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. J. Comput. Phys. 1977, 23, 327–341. (30) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98, 10089–10092. (31) Heinig, M.; Frishman, D. Nucleic. Acids. Res. 2004, 32, W500–W502. (32) Roccatano, D.; Colombo, G.; Fioroni, M.; Mark, A. E. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12179–12184. (33) Yamazaki, K.; Iwura, T.; Ishikawa, R.; Ozaki, Y. J. Biochem. 2006, 140, 49–56. (34) Searle, M. S.; Zerella, R.; Williams, D. H.; Packman, L. C. Protein Eng. 1996, 9, 559–565. (35) Cierpicki, T.; Otlewski, J. J. Biomol. NMR 2001, 21, 249–261. (36) Rothemund, S.; Weisshoff, H.; Beyermann, M.; Krause, E.; Bienert, M.; M€ugge, C.; Sykes, B. D.; S€onnichsen, F. D. J. Biomol. NMR 1996, 8, 93–97.
6660
dx.doi.org/10.1021/jp111448a |J. Phys. Chem. B 2011, 115, 6653–6660
