|Home | About | Journals | Submit | Contact Us | Français|
We used CD, UV resonance Raman spectroscopy, and molecular dynamics simulation to examine the impact of salts on the conformational equilibria and the Ramachandran Ψ angle (un)folding Gibbs free energy landscape coordinate of a mainly polyalanine α-helical peptide, AP of sequence AAAAA-(AAARA)3A. NaClO4 stabilizes α-helical-like conformations more than does NaCl, which stabilizes more than Na2SO4 at identical ionic strengths. This α-helix stabilization ordering is the reverse of the Hofmeister series of anions in their ability to disorder water hydrogen bonding. Much of the NaClO4 α-helix stabilization results from ClO4− association with the AP terminal –NH3+ groups and Arg side chains. ClO4− stabilizes 310-helix conformations but destabilizes turn conformations. The decreased Cl− and SO42− AP α-helix stabilization probably results from a decreased association with the Arg and terminal –NH3+ groups. Cl− is expected to have a smaller binding affinity and thus stabilizes α-helical conformations intermediately between NaClO4 and Na2SO4. Electrostatic screening stabilizes π-bulge conformations.
The mechanism(s) whereby peptides and proteins fold into their native states are poorly understand (1–6). The well-known Levinthal paradox (7) clearly demonstrates that proteins do not fold through a random search of their conformational space since this would take longer than the age of our universe. Recent energy landscape models (1, 3, 8, 9) propose that funnel-shaped folding energy landscapes occur, where the native state is accessed via a strategically sloped energy landscape that funnels unfolded conformations toward the native folded state (3, 10, 11).
In the work here we use CD, UV resonance Raman spectroscopy (UVRR), and molecular dynamics simulations to examine the Gibbs free energy landscape along the Ψ Ramachandran angle folding coordinate of a mainly polyalanine peptide, AP of sequence AAAAA(AAARA)3A, in pure water and in the presence of NaClO4, NaCl, and Na2SO4. AP-like peptides have been the subject of intensive experimental (12–30) and theoretical (31–42) studies which have probed the mechanism(s) of α-helix folding and unfolding. The AP peptide is ~50%α-helical-like at 0 °C and melts to PPII-like conformations at higher temperatures (43–47). We previously found that AP (un)folding is not a simple two-state process because it involves other secondary structure conformations such as π-bulge and 310-helix and turn structures (22, 31, 32, 41, 42, 48, 49). We examined the dependence of the AP conformational equilibrium and melting on the presence of different salts and find that ion binding and electrostatic screening significantly modulate the Gibbs free energy landscape and stabilize α-helix conformations. We also find that the ordering of salt stabilization of the α-helical content can be explained by Collins et al. (law of matching water affinities) (50–52).
The 21-residue peptide AP of sequence AAAAA(AAARA)3A was purchased from AnaSpec Inc. (>95% purity). Anhydrous NaCl, NaClO4, and Na2SO4 were purchased from J. T. Baker (>99% purity). All AP samples were prepared at 1.0 mg/mL concentrations at pH 7.
The CD spectra were measured by using a Jasco-715 spectro-polarimeter, using a 200 μm path length cuvette. We coadded ten individual spectra.
The UVRR spectrometer was described in detail by Bykov et al. (53). Briefly, 204 nm UV light was obtained by generating the fifth anti-Stokes Raman harmonic of the third harmonic of a Nd:YAG laser (Coherent, Infinity). We used a spectral accumulation time of 5 min for each measurement and coadded four accumulations.
We performed replica exchange molecular dynamics (REMD) studies of AP immersed in pure water and in a 0.2 M NaClO4 aqueous solution. The temperatures studied range from 270 to 505 K. The simulation details are given in the Supporting Information. To investigate the mechanisms which govern helix stabilization, we used Chimera (54) to calculate the average ion occupancy surrounding the AP peptide.
Figure 1 shows the temperature dependence of the CD spectra of AP in pure water. The lower temperature CD spectra show two troughs at 222 and 206 nm that are characteristic of α-helix conformations (55). As the temperature increases, the ellipticity at 222 nm, Θ222, becomes less negative, indicating α-helix melting. The isosbestic point at 202 nm indicates that this melting appears spectroscopically as a “two-state” process. Previous work by our group demonstrated that the AP α-helix conformation melts to a dominantly PII-like conformation (56).
Addition of NaClO4 (Figure 2a) increases the AP α-helical content at all temperatures, as evident from the more negative values of Θ222; however, the CD changes are relatively small between 1 and 2MNaClO4 concentrations. As discussed below, Tm also increases as the NaClO4 concentration increases.
Panels b and c of Figure 2 show similar AP melting curves for NaCl and Na2SO4. Again, the α-helix fractions increase as evident from the more negative Θ222 values as the salt concentrations increase. Again, little increase occurs between 1 and 2M salt concentrations.
To quantitatively model α-helix melting, we calculated the α-helical conformational fraction, fα, using a two-state model (eq 1) by utilizing the reported Θ222 values for the pure α-helix ([θ]α=−26000 deg · cm2 · dmol−1) and the “pure melted” conformations ([θ]r = −3500 deg · cm2 · dmol−1) (57).
As shown below, the structure is actually more complex than a two-state transition, but significant useful thermodynamics information associated with the resulting quasi-two-state transition can be extracted from the CD data.
We also calculated the melting thermodynamic parameters by fitting the calculated equilibrium α-helix fraction, fα to
The fits to eq 2 versus T−1 are highly linear with R2 values of >0.995. Table 1 lists the calculated values for ΔH, ΔS, and the resulting estimated Tm values. Figure 3 shows that Tm increases for all salts as their concentrations increase (except at the highest Na2SO4 concentration). NaClO4 and NaCl stabilize α-helices more than does Na2SO4. Tm increases as the Na2SO4 concentration increases to 0.333M(ionic strength of 1.0 M) but then begins to decrease at a concentration of ~0.667 M (ionic strength of 2.0 M).
ΔH becomes more negative as the NaClO4 concentration increases to 0.2 M but then becomes less negative at higher NaClO4 concentrations. A similar trend occurs for ΔS. Thus, the α-helix becomes more (less) favored enthalpically (entropically) as the salt concentration increases. ΔG, the difference between ΔH and TΔS, for α-helix formation becomes more negative as the NaClO4 concentration increases.
In contrast, ΔH and ΔS both become more negative as the NaCl concentration increases to 1.0 M, but the values saturate upon concentration increases to 2.0 M.
Na2SO4 also stabilizes the α-helix giving rise to an increasingly negative ΔG. However, ΔH and ΔS only show a modest dependence on the Na2SO4 concentration. They both change together to make ΔG increasingly negative as evident from the Figure 2c melting data for the lower concentrations.
Figure 4 shows the temperature dependence of the AP 204 nm UVRR spectra. The AmI band (~1660 cm−1) results from a mainly C=O stretching vibration. The AmII band (~1550 cm−1) derives from out-of-phase motion of C–N stretching and N–H bending. The Cα–H doublet (~1372 and ~1393 cm−1) derives from a Cα–H bending vibration which is resonance enhanced because of coupling of Cα–H bending to N–H bending. The intensities of the Cα–H bending bands increase as the concentration of nonhelical conformations increases (58). The AmIII bands arise from vibrations which involve in-phase contributions of C–N stretching and N–H bending. The AmIII region contains three subbands: the AmIII1 band (~1336 cm−1), the AmIII2 band (~1306 cm−1), and the AmIII3 band (~1250 cm−1).
The AmIII3 band is the most conformationally sensitive because it involves Ψ-angle-dependent coupling between N–H bending and the Cα–H bending motions (59). For example, the AmIII3 band of the α-helix appears at 1258 cm−1 and contains little Cα–H bending. However, it shifts to 1247 cm−1 in the PPII conformation (60) and contains significant Cα–H bending. The α-helixAmIII3 band cross section is roughly half that of the PPII-like conformation because of the α-helix conformation electronic transition hypochromism (61). As the temperature increases, the intensities of the Cα–H bands increase, indicating α-helix melting. Also, the AmIII3 band frequency shifts from ~1258 cm−1 at 2 °C to ~1247 cm−1 at 80 °C.
The UVRR of AP in 0.2 M NaCl, 0.2 M NaClO4, and 0.0667 M Na2SO4 solutions all indicate α-helix melting as the temperature increases (spectra not shown). The spectra of samples without ClO4− or SO42− were normalized to the AmI band integrated intensity, which shows little variation upon peptide conformational changes (62).
To calculate the α-helical fractions, we calculated the temperature-dependent basis spectra of the PPII-like conformation by using the method of Lednev et al. (57). We then digitally smoothed and then subtracted the appropriate amount of the PPII-like conformation basis spectra from the measured and smoothed UVRR of AP.
The relative amount of the PPII conformation subtracted is the maximum amount of the PPII basis spectrum which minimized the Cα–H region intensity, with the constraint that no negative features occur in the difference spectrum. The basis spectral intensities subtracted are directly proportional to the concentration of the PPII conformation at each temperature (57). The resulting difference spectra should result only from non-PPII conformations and appear to be mainly α-helix-like.
Figure 5, which shows the temperature dependence of the non-PPII fraction, indicates that NaClO4 is the most “helix”-stabilizing salt, followed by NaCl, Na2SO4, and then pure water. These results agree with the CD results above and with the salt ordering previously observed by others (63).
Figure 6 shows the temperature dependence of the calculated non-PPII, α-helix-like spectra of AP in NaClO4, NaCl, and Na2SO4 and the difference spectra between the different salt solution spectra. At all temperatures we observe a triplet of bands which are the hallmark of α-helix-like UVRR spectra. The 30 °C spectrum shows a change in the AmIII3 band shape as earlier noted by Mikhonin and Asher (48). The 30 °C AmIII3 band slightly narrows, while the maximum becomes more sharply peaked. This band shape change, which appeared as a more simple band narrowing in our previous poorer S/N spectra, was ascribed to a decrease in the concentrations of 310-helix and π-bulge conformations relative to the α-helix concentration as the temperature increases.
The most prominent difference between the different salt spectra is that an ~1200 cm−1 band that occurs in pure water, in NaCl, and in Na2SO4 disappears in NaClO4 as indicated by the ~1200 cm−1 troughs in the difference spectra. A new band occurs at ~1240 cm−1 in NaClO4 solution, as evident by the trough in the difference spectra between the NaClO4 solution and the other salts and pure water (most clearly at temperatures below 20 °C). Previous work (48, 62) indicated that an AmIII3 band at ~1200 cm−1 derives from turn structures, while the AmIII3 band at ~1240 cm−1 derives from 310-helix conformations. Therefore, we conclude that NaClO4 selectively stabilizes 310-helix conformations which are replaced in pure water, in NaCl, and in Na2SO4 by turn conformations.
We calculated the Gibbs free energy landscapes of AP along the Ψ-folding coordinate from the UVRR (Figures 7 and and8)8) by using the methodology of Mikhonin et al. (48, 53, 60, 64). We calculate the Ψ angle probability distribution from the AmIII3 band shape and utilize the Boltzmann relationship to calculate the Gibbs free energy landscape.
The calculated AP Gibbs free energy landscapes in pure water and in 0.2 M NaClO4 at 2.4 °C (Figure 7) show a broad α-helix-like conformational region which includes Ψ angles corresponding to 310-helices and π-bulges and the broad PPII region. The presence of 0.2 M NaClO4 selectively decreases the Gibbs free energies of the 310-helix conformations but increases the free energy of the turn conformations. The α-helix-like region, the turn region, and the PPII region are separated by high activation barriers due to steric clashes. These results are consistent with the calculated Φ and Ψ dependence of the Gibbs free energies of peptide conformations (65). High activation energies are expected between α-helix and turn conformations and PPII conformations. Our ability to monitor turn conformations is important for insight into the mechanisms of α-helix melting since these turn conformations are likely to serve as intermediates along the reaction coordinate that links α-helix-like to the melted PPII-like conformations (66, 67).
The energy landscape (Figure 8) is bumpy within the α-helix-like basin. Within this basin the pure α-helix conformation (Ψ ~ −45°) is always lowest in energy, followed by the π-bulge conformation (Ψ ~ −70°). The 310-helix conformation (Ψ ~ −20°) lies at a slightly higher energy. Both the π-bulge and 310-helix conformations appear to show activation barriers between their minima and that of the α-helix conformation. The relative energy of the π-bulge conformation compared to the α-helix conformation is highest in pure water. As the temperature increases, the α-helix basin Gibbs free energy increases, indicating that the α-helix is destabilized relative to the PPII conformation. We have drawn the energy landscape as a projection onto the Ψ angle coordinate. Traversing the energy landscape from one conformation to another could involve complex dynamics and involve significant dynamics involving Φ angle excursions and excursions in other coordinates.
Figure 8 also shows the dependence of the conformational energies as a function of the salts dissolved in the AP solution at 0.2 M ionic strengths. For all temperatures, the lowest α-helix Gibbs free energies occur in the presence ofNaClO4, followed by NaCl, Na2SO4, and pure water. The 310-helix conformation is selectively stabilized by NaClO4.
The NaClO4 α-helix stabilization was qualitatively reproduced by REMD simulations. Figure 9 shows the theoretically calculated AP non-PPII fraction as a function of temperature in 0.2 M NaClO4 solution and in pure water (all conformations with dihedral angles Ψ < +50° are counted as non-PPII conformations). The AP α-helical-like conformation (non-PPII conformations) concentrations in NaClO4 are greater than in pure water at all temperatures.
To investigate the mechanisms governing the helical stabilization, we studied the equilibrium ClO4− distribution around AP. The average ClO4− concentrations around AP were calculated by using Chimera (54), where we calculated the average occupancy of ClO4− within a 3D grid over the trajectory frames. Figure 10 shows the concentration distribution where a green volume indicates a region of increased ClO4− occupancy.
The largest increased ClO4− occupancy occurs around the N-terminus. Panels a and b of Figure 10 also show increased ClO4− occupancies around Arg 9 and Arg 14, with no increased ClO4− occupancy around Arg 19, presumably due to the neutralization of Arg 19 charge by the carboxylate. Figure 10c shows that the ClO4− occupancy increases in the region between the Arg side chain and the peptide backbone.
Panels a and b of Figure 11, which show the radial distribution functions for ClO4−, for the average β-carbon of Ala, the terminal –NH3+, and the three different Arg ζ-carbons, indicate that ClO4− is located on the average 4.5A ° from the Ala β carbon and ~4 A° from the terminal –NH3+ nitrogen, while it is on the average ~5 A° from the Arg ζ carbon. The distribution of ClO4− around the terminal –NH3+ is highest, indicating that the largest association of ClO4− with AP occurs around the terminal –NH3+. Arg 9 has the highest affinity for ClO4−, followed by Arg 14 and then Arg 19.
ClO4− binding to the terminal –NH3+ and to Arg is expected from Collion’s matching water affinity model (50–52) where ion pair formation occurs preferentially between oppositely charged ions of similar charge densities. In this model small ions of high charge density tend to preferentially form ion pairs. Large ions of low charge density also preferentially form ion pairs. Ion pairing between oppositely charged ions of high and low charge density is less favorable. Thus, for AP we expect that the weakly hydrated (with low charge density) N-terminal and Arg side chains (50–52) will most strongly bind to the weakly hydrated, low charge density ClO4− which will promote neutralization of the electrostatic interactions within AP which will significantly stabilize the α-helix conformation.
Both the CD and UVRR results show that salts stabilize α-helical-like conformations of AP, with efficiencies: NaClO4 > NaCl ~Na2SO4. Numerous previous studies (68, 69) have proposed that these salts affect the protein/peptide stability through three main effects.
Thus, for AP we expect that the weakly hydrated N-terminal and Arg side chains (50–52) most strongly bind to the weakly hydrated chaotrope ClO4−. The resulting charge neutralization will decrease electrostatic interactions within AP which should significantly stabilize the α-helix conformation. Our molecular dynamics simulations directly observe association of ClO4− with the terminalNH3+ group and the Arg. SO42− is expected to show the least ion pairing with NH3+ and Arg, while Cl− should show intermediate ion pairing. These ion pairing propensities predict that NaClO4 should be the most helix stabilizing, followed by NaCl and then Na2SO4.
Table 2 summarizes the expected impact of ions on AP α-helix stability from different effects. Our observations show that at identical ionic strengths α-helical-like conformations are most stabilized by NaClO4, followed by NaCl, then Na2SO4, and then pure water. For NaClO4 to exhibit the most stabilization, the positive impact from the specific ion binding effect must override the negative impact of the Hofmeister effect. The fact that NaCl stabilizes α-helical-like conformations less than does NaClO4 but more than Na2SO4 suggests that the Hofmeister effect (water structure modification) has little impact.
The fact that π-helix (bulge) is disfavored in pure water but is stabilized in the presence of these three salts demonstrates that importance of electrostatic screening, where formation of the π-bulge must overcome repulsion between Arg side chains spaced at i, i + 5 positions. In contrast, the Arg side chains are spaced further apart for the α-helix and the 310-helix. We are continuing to study why NaClO4 stabilizes the 310-helix but destabilizes the turn structure.
We used CD and UV resonance Raman spectroscopy and molecular dyamcis to study the solution conformation of a mainly polyalanine peptide containing Arg groups for solubility. We calculated the Gibbs free energy landscape along the Ramachandran Ψ angle folding coordinate. We observe that at identical ionic strengths α-helical-like conformations are stabilized most byNaClO4 due to preferential ion binding of ClO4− to the terminal NH3+ and Arg side chains. ClO4− stabilizes 310-helices but destabilizes turn conformations. Cl− has a smaller binding affinity and thus stabilizes α-helical conformations intermediate between NaClO4 and Na2SO4. Electrostatic screening stabilizes π-bulge conformations. We find that we can understand ion association to the peptide through the Collins “laws of matching water affinities model”.
We thank Aleksandr V. Mikhonin, Zeeshan Ahmed, Bhavya Sharma, Lu Ma, and Zhenmin Hong for useful discussions.