Transition state and ground state properties of the helix–coil transition in peptides deduced from high-pressure studies

Significance

High pressure unfolds proteins due to a larger volume of the native state compared with the unfolded state, but the molecular origin of the volume changes is still under debate. Virtually no information is available on volume changes associated with secondary structure formation. We reveal that α-helices become stabilized with increasing pressure, which explains the commonly observed residual helical structure in pressure-unfolded proteins. The transition state for the helix–coil transition has a larger volume than both the helical and the coil state. Thus, addition or removal of a helical residue proceeds through a transitory high-energy state with a large volume, possibly due to the presence of unsatisfied hydrogen bonds, although steric effects may also be involved.

Abstract

Volume changes associated with protein folding reactions contain valuable information about the folding mechanism and the nature of the transition state. However, meaningful interpretation of such data requires that overall volume changes be deconvoluted into individual contributions from different structural components. Here we focus on one type of structural element, the α-helix, and measure triplet–triplet energy transfer at high pressure to determine volume changes associated with the helix–coil transition. Our results reveal that the volume of a 21-amino-acid alanine-based peptide shrinks upon helix formation. Thus, helices, in contrast with native proteins, become more stable with increasing pressure, explaining the frequently observed helical structures in pressure-unfolded proteins. Both helix folding and unfolding become slower with increasing pressure. The volume changes associated with the addition of a single helical residue to a preexisting helix were obtained by comparing the experimental results with Monte Carlo simulations based on a kinetic linear Ising model. The reaction volume for adding a single residue to a helix is small and negative (−0.23 cm³ per mol = −0.38 ų per molecule) implying that intrahelical hydrogen bonds have a smaller volume than peptide-water hydrogen bonds. In contrast, the transition state has a larger volume than either the helical or the coil state, with activation volumes of 2.2 cm³/mol (3.7 ų per molecule) for adding and 2.4 cm³/mol (4.0 ų per molecule) for removing one residue. Thus, addition or removal of a helical residue proceeds through a transitory high-energy state with a large volume, possibly due to the presence of unsatisfied hydrogen bonds, although steric effects may also contribute.

Understanding the effect of pressure on protein stability and dynamics provides insight into fundamental principles and mechanisms of protein folding (1). For most proteins, increasing pressure shifts the folding equilibrium toward the unfolded state, which, according to Le Chatelier’s principle, shows that the native state has a larger volume than the unfolded state (2–4). The origin of the volume increase upon folding has been discussed controversially for a long time. The major obstacle in the interpretation of volume changes is opposing contributions from different effects. Analysis of high-resolution X-ray structures suggested that formation of intramolecular hydrogen bonds and van der Waals interactions in the native state lead to a decrease in atomic volumes and thus to a decrease in protein volume upon folding (5). Further, water around hydrophobic groups has a larger volume than bulk water, which also leads to a decrease in volume upon burial of hydrophobic groups in the native protein (6). On the other hand, formation of ordered water structures around charged groups (7) and solvation of the peptide backbone decrease the water volume (6), which leads to a volume increase upon folding. Recent experimental results suggested that volume changes associated with transfer of groups from solvent to the protein interior upon folding are small and that the formation of void volumes in native proteins is the major origin of the volume increase upon folding (8).

Volume changes associated with the formation of protein folding transition states are only poorly characterized. High-pressure stopped-flow experiments on tendamistat (9) and cold shock protein (10), as well as pressure-jump experiments on cold-shock protein (10) and an ankyrin repeat domain (11), revealed that volumes of protein folding transition states are close to the volume of the native state and may even exceed the native state volume under some conditions (9–11). Similar results were found for the folding and unfolding reactions of ribonuclease A (12), which are both complex and dominated by kinetic coupling between slow prolyl isomerization and protein folding reactions (13–16). All folding and unfolding reactions of RNase A reveal positive activation volumes, indicating that the transition state volumes are larger than those of the native and unfolded state (12).

To understand the origin of volume changes associated with protein folding it is important to dissect contributions of intramolecular interactions, such as secondary structure formation, from contributions of packing deficiencies in the native state. Volume changes associated with the formation of protein α-helices can be obtained by studying the effect of pressure on stability and folding–unfolding dynamics of α-helical peptides. Alanine-based peptides form helical structures with intramolecular hydrogen bonds and solvent-accessible side chains and are thus well-suited to study the effect of pressure on secondary structure formation in the absence of void volumes formed in the protein core of native proteins. The effect of pressure on the stability of alanine-based helical peptides has been addressed both experimentally and in simulations albeit with different conclusions. FTIR experiments revealed an increase in helix stability with increasing pressure, indicating a smaller volume of the helical state compared with the unfolded state (17). Molecular dynamics simulations, in contrast, suggested that the unfolded state of Ala-based peptides has a slightly smaller volume than the helical state (18). The simulations further revealed a change in geometry and length of hydrogen bonds with increasing pressure, which should have an effect on FTIR bands and NMR chemical shifts. Changes in hydrogen bond length in protein secondary structures with increasing pressure were experimentally confirmed in high-pressure NMR studies on ubiquitin (19).

To determine the effect of pressure on stability and dynamics of α-helices in the absence of tertiary structure and without contributions from changes in spectroscopic properties, we measured triplet–triplet energy transfer (TTET) at various pressures in an Ala-based 21-amino-acid helical peptide. TTET coupled to a folding–unfolding equilibrium yields information on local conformational stability and dynamics on the nanoseconds to microseconds time scale (20–22). Intrachain TTET in polypeptide chains occurs through loop formation and is based on van der Waals contact between a triplet donor and a triplet acceptor group (23, 24). If the triplet labels are introduced in proteins and peptides at sites that are not in contact in the folded structure, unfolding in the region between the labels has to occur before TTET. The resulting TTET kinetics yield information on the local folding–unfolding dynamics and on the local stability in the region between the TTET labels (20–22). In our previous work we introduced the triplet donor xanthone (Xan) and the triplet acceptor naphthylalanine (Nal) into α-helical peptides with i,i+6 spacing, which places them at opposite sides of the helix and prevents TTET in the helical state (Fig. 1). Thus, at least partial unfolding of the helix in the region between the labels is required before TTET can occur (20, 22). The overall reaction can be described by the three-state model shown in Fig. 1. Because the folding–unfolding dynamics and loop formation occur on a similar time scale (20, 22) and both the folded and the unfolded state are populated to significant amounts in equilibrium, the two observable rate constants λ₁ and λ₂ for TTET and their corresponding amplitudes A₁ and A₂ yield the rate constants for local helix formation and unfolding between the labels k_f and k_u, as well as the rate constant for loop formation (k_c) (SI Text). In addition, the local equilibrium constant (K_eq) for helix stability in the region between the labels can be obtained from K_eq = k_f/k_u. TTET experiments do not require perturbation of the helix–coil equilibrium and thus yield information on equilibrium fluctuations of the helix. We have previously investigated local stability and dynamics in various regions of the 21-amino-acid Ala-based helical peptide, which showed a position-independent helix formation rate constant (k_f) and a position-dependent helix unfolding rate constant (k_u) with faster unfolding at the termini compared with the center (20). This results in higher local helix stability in the center compared with the ends, which is expected from helix–coil theory based on a linear Ising model and in agreement with previous results from hydrogen–deuterium exchange experiments (25).

Fig. 1.

Schematic representation of TTET coupled to a helix–coil equilibrium. The equilibrium between a folded (F) and an unfolded or partially unfolded (U) conformation between the labels is monitored by fast and irreversible TTET through loop formation in the ensemble of unfolded conformations (U*). The triplet labels xanthonic acid (Xan, blue) and 1-naphthylalanine (Nal, red) are placed on opposite sides in the central region of the helix with i,i+6 spacing.

Because TTET coupled to the helix–coil transition gives information on local equilibrium constants, independent of changes in spectroscopic properties of the helix, this method is perfectly suited to study the effect of pressure on local helix stability. In addition, pressure-dependent TTET experiments yield the activation volumes for helix folding and unfolding. Here, we investigate the effect of pressures on stability and dynamics of the α-helix–coil transition in the center of the 21-amino-acid Ala-based peptide using a high-pressure laserflash setup. The results reveal that the helical structure becomes stabilized with increasing pressure indicating a volume decrease upon helix formation. The dynamics of helix formation and unfolding both slow down with increasing pressure, which shows that the transition state for growth and shrinking of the helical structure has a larger volume than both helical and unfolded state and points at a non-hydrogen-bonded transition state structure.

Results and Discussion

Effect of Pressure on Local Helix Stability and Dynamics.

To investigate the effect of pressure on local helix dynamics and stability, we used a 21-amino-acid Ala-based model peptide with arginine residues placed with i,i+5 spacing. The triplet pair Xan and Nal was introduced in the central region of the peptide at positions 7 and 13, respectively, which yields the peptide sequence:

Xan was introduced by coupling 9-oxoxanthen-2-carboxylic acid to the side chain of the nonnatural amino acid α,β-diaminopropionic acid. The nonnatural amino acid 1-naphthylalanine was directly incorporated into the peptide during solid-phase peptide synthesis. The C terminus was amidated but, in contrast with our previous work, the helical peptide studied here has a free N terminus, which decreases helix stability and leads to about 50% helical conformations in the center of the peptide at ambient pressure (22). The similar populations of unfolded and helical conformations provide a high sensitivity of the measurements for both an increase and a decrease in helix stability, because it was not clear whether pressure stabilizes or destabilizes the helical structure.

We investigated the effect of pressure on helix stability in the center of the peptide by TTET measurements in the range between 0.1 and 390 MPa (1–3,900 bar). The experiments were initiated by a 4-ns laserflash at 355 nm, which produces the xanthone triplet state that can be detected by its strong absorbance band at 590 nm (24, 26). TTET to naphthylalanine leads to a decay in xanthone triplet absorbance band (Fig. 2A). The triplet decay curves can be described by the sum of two exponentials under all conditions with the rate constants λ₁ and λ₂ (Fig. 2B) and their respective amplitudes A₁ and A₂ (Fig. 2C). This result indicates that both the unfolded and the helical conformation are populated to significant amounts in equilibrium at all pressures (20). Fig. 2A shows that TTET kinetics become slower with increasing pressure. The results from the double-exponential fits reveal that increasing pressure decreases both λ₁ and λ₂ and leads to a gain in amplitude of the slower kinetic phase at the expense of the amplitude of the faster phase (Fig. 2).

Fig. 2.

Pressure dependence of helix–coil dynamics in the center of a 21-amino-acid helical peptide measured by TTET. (A) Xanthone triplet decay at the indicated pressures monitored by the absorbance change at 590 nm. All curves were fitted by the sum of two exponentials. Additionally, a minor slow kinetic phase is observed in the triplet decay curves, which corresponds to the intrinsic triplet lifetime of xanthone and probably reflects a small fraction of oligomeric molecules. (B) Pressure dependence of the two observable rate constants λ₁ and λ₂ and of their respective amplitudes, A₁ and A₂ (C). The black lines in B and C reveal the results for λ₁ and λ₂, and their amplitudes from global fits of all kinetic traces according to the analytical solution of the three-state model shown in Fig. 1 (SI Text). Global fitting of all kinetic traces additionally yields the pressure dependence of the microscopic rate constants, k_u, k_f, and k_c as indicated in B, which results in activation volumes of = 3.3 ± 0.3 cm³/mol, = 1.7 ± 0.5 cm³/mol, and = 3.6 ± 0.4 cm³/mol.

The effect of pressure on the rate constant of a reaction depends on the activation volume (ΔV^0‡) according to

However, the two rate constants observed for TTET (λ₁ and λ₂) do not directly correspond to folding or unfolding rate constants in Fig. 1 but are functions of k_f, k_u, and k_c (Fig. 1), which can be determined by fitting the solution of the three-state mechanism (SI Text) to the observable rate constants (Fig. 2B) and their amplitudes (Fig. 2C). The quality of the fit improves significantly when the kinetic traces measured at the different pressures are fitted globally assuming a linear pressure dependence of ln(k_f), ln(k_u), and ln(k_c) according to Eq. 1. The results from the global fit reveal that all rate constants decrease with increasing pressure with activation volumes of 1.7 ± 0.5 cm³/mol, 3.3 ± 0.3 cm³/mol, and 3.6 ± 0.5 cm³/mol for , , and , respectively (Figs. 2B and 3B). This result indicates that the transition state for helix formation and unfolding has a larger volume than both the helical state and the unfolded state. In addition, loop formation in the coil state (k_c) slows down with increasing pressure.

Fig. 3.

Pressure dependence of (A) the equilibrium constant for helix formation (K_eq = k_f/k_u) in the peptide center and (B) the rate constants for helix formation (k_f) and unfolding (k_u). The lines represent the results from global fits to the experimental data (Fig. 2B) resulting in a reaction volume of −1.6 cm³/mol. The filled circles are the results from the simulations based on the linear Ising model with the parameters given in Eqs. 10–12. The quality of the results was improved by performing simulations up to s-values corresponding to p = 2,000 MPa (3).

The equilibrium constants at the various pressures can be obtained from K_eq = k_f/k_u. Applying the Planck equation (Eq. 2)

yields the reaction volume for helix formation in the region between the TTET labels, which equals the difference between the activation volumes for helix formation and unfolding (Eq. 3):

The equilibrium constant for helix formation increases with increasing pressure, and the fit according to Eq. 2 gives a negative reaction volume (ΔV⁰) of −1.6 ± 0.6 cm³/mol (Fig. 3A). This result shows that the helical state has a smaller volume than the unfolded state and thus becomes stabilized with increasing pressure.

Dissociation of phosphate has a reaction volume of about −25 mL/mol, which leads to a decreases in its pK_a value of about 0.3 units per 100 MPa and results in a decrease in pH of about 1.2 units in the pressure range applied in our TTET experiments (27). The Ala-based helical peptide used in this study contains three arginine residues and a free N terminus, which have pK_a values of 12.2 and 9.6, respectively. These values should be independent of pressure because the deprotonation reaction of neither Arg nor the N terminus leads to a change in the number of charged species (27). The pressure-induced decrease in pH from 7.0 to 5.8 between 0.1and 390 MPa applied in our experiments therefore does not lead to a change in the protonation state of the helical peptide and should not influence the helix–coil transition. This was confirmed by measuring TTET kinetics in the pH range between 5.8 and 7.0, which are insensitive to pH (Fig. S1). In addition, pressure in the applied range does not influence the triplet donor lifetime (Fig. S2).

Effect of Pressure on the Elementary Rate Constants for Helix Elongation and Shrinking.

The activation and reaction volumes obtained from the TTET experiments are apparent volume changes for helix formation and closure between the labels that cannot be directly converted to the volume changes for addition or removal of a single helical residue, because the helix–coil transition is not a two-state process (20, 28–33). The dynamics and stability in the helix center are rather influenced by the coupling of many microscopic opening and closing events at the different positions in the peptide. Our previous studies showed that a kinetic linear Ising model with 1D boundary diffusion mechanism is quantitatively able to describe the experimentally determined position dependence (20) and length dependence (22) of dynamics and stability of the helix–coil transition. Applied to the helical peptide used in our study, the helix is represented by a finite sequence of 21 identical residues, which can be either in the helical state h or the coil state c, leading to the following type of conformations:

The following equations describe the dynamic and equilibrium properties of the system (32, 34):

helix elongation:

helix nucleation:

coil nucleation:

Although this model only accounts for nearest-neighbor interactions (34), it is able to quantitatively describe experimental data on the helix–coil transition in a variety of different peptides (28, 31, 35). We further used the kinetic extension of the model introduced by Schwarz (32), which introduces additional factors for helix nucleation γ_h and coil nucleation γ_c to account for kinetic effects on the nucleation reactions relative to elongation:

We performed Monte Carlo simulations based on Eqs. 4–9 and compared the results with the experimental data to obtain information on the activation volumes of the elementary rate constants for helix growth () and shrinking () and on the reaction volume for adding a single helical residue at the helix–coil border (; see Eq. 5). In all simulations k₁ was set to 1, the nucleation parameter σ was 0.003 and the kinetic parameters were γ_c = γ_h = 2 (20, 22). As the model peptide has a destabilized N-capping region, we distinguish between the average s-value of the positions 5–21 and the s-value for the N-terminal cap assigned to positions 2 to 4 (22). The folding and unfolding dynamics were simulated for different s-values (Fig. 4A) and analyzed by computing first-passage times (FPTs) (20, 22). Because the experiments probe a region of five residues between the labels, we analyzed the results from the simulations by monitoring the dynamics of the whole segment between the labels instead of single-site fluctuations. The segment of five residues between the labels was counted as helical if at least four of its residues were helical at a certain time (20, 22). In addition, the starting conformations had to contain at least one helical residue. The distribution of the FPTs was fitted by single-exponential functions yielding the reduced rate constants of local helix folding and unfolding k_f/k₁ and k_u/k₁ for each s-value (Fig. 4B). Because the simulations exhibit a considerable statistical error, especially for small s-values, they were performed up to s-values corresponding to P = 2,000 MPa (Fig. S3), which gave more reliable results for the pressure dependence of the kinetic and equilibrium parameters in the experimentally accessible region between 0.1 and 390 MPa (SI Text). The best agreement between simulated and experimental data (Fig. 3A) is obtained with a pressure-independent s-value for the N-capping region of 0.099 ± 0.004 and a pressure-dependent general s-value that can be described by (Fig. 4A)

According to Eq. 2 this corresponds to s(0.1 MPa) = 1.04 and (−0.38 ų per molecule) for addition of a single helical residue.

Fig. 4.

Pressure dependence of (A) the s-value (Eq. 5) and (B) the reduced rate constants for local helix folding and unfolding, k_f/k₁ and k_u/k₁, respectively, obtained from the Monte Carlo simulations. The solid circles in A represent the results from the simulations that describe the experimental data (Fig. 3A) and the line represents the fit with the results given in Eq. 10 resulting in =−0.23 cm³/mol. (C) Pressure dependence of the rate constants for adding (k₁) and removing (k_-1), a single helical residue to–from an existing helix (Eq. 5), corresponding to activation volumes of = 2.2 cm³/mol and = 2.4 cm³/mol. The quality of the results was improved by performing simulations up to s-values corresponding to p = 2,000 MPa (Fig. S3).

The simulations further yield reduced rate constants k_f/k₁ and k_u/k₁ for helix formation and unfolding in the helix center (Fig. 4B). Comparison of the reduced rate constants, with the experimentally determined rate constants for helix folding and unfolding in the region between the labels, k_f and k_u (Fig. 2C), yields a scaling factor for k₁, which gives the actual rate constant for addition of a single helical residue (Eq. 5). The experimentally determined rate constants for folding and unfolding are described by the simulations (Fig. 3B) with a pressure-dependent scaling factor (Fig. 4C) that is described by

which corresponds to an elementary rate constant for adding a helical residue at 0.1 MPa of k₁(0.1 MPa) = 8.7 × 10⁶ s⁻¹ and = 2.2 cm³/mol (3.7 ų per molecule) The pressure dependence of k_-1 is obtained from k₁ in combination with the s-value (Eq. 5), which yields (Fig. 4C)

corresponding to k_-1(0.1 MPa) = 8.4 × 10⁶ s⁻¹ and = 2.4 cm³/mol (4.0 ų per molecule).

These results reveal that the reaction volume for adding a single helical residue is small and negative, whereas the activation volumes for adding or removing a helical residue are both large and positive. When a residue is added to a helix, backbone–water, hydrogen bonds and nonspecific intramolecular hydrogen bonds are lost (Fig. 5). Concomitantly, an intramolecular i,i+4 backbone hydrogen bond is formed and additionally, the peptide backbone becomes partially shielded from solvent. However, this solvent-shielding contribution is expected to be relatively minor for alanine-based peptides, which permit water to solvate the intrahelical hydrogen bonds due to the small size of the Ala side chains (36). The negative reaction volume for adding a helical residue suggests that intrahelical i,i+4 hydrogen bonds have a slightly smaller volume than backbone–water and long-range backbone–backbone hydrogen bonds in the unfolded state. Contributions from void volumes to the reaction volume are probably small, because the α-helix is a close-packed structure. However, we cannot exclude the possibility that small void volumes accompany helix formation but are compensated by the volume decrease upon formation of intramolecular hydrogen bonds.

Fig. 5.

Schematic representation of the structural and volume changes during addition of a single helical residue to an existing α-helix. The coil state (hcc) has the carbonyl oxygen and amide proton hydrogen bonded to water molecule. Upon adding a helical residue (hhc) an intramolecular hydrogen bond is formed and water is released. The large volume of the transition state (‡) indicates that neither an intramolecular hydrogen bond nor peptide–water hydrogen bonds are formed.

The larger volume of the transition state compared with the volumes of either the helical or the coil state shows that formation or removal of a single helical residue proceeds through a high-energy state with a large volume. A plausible explanation for this large transition state volume is the presence of amide groups that are neither satisfied by hydrogen bonds to water nor by intramolecular hydrogen bonds (Fig. 5). In this case, the loss of a hydrogen bond between a carbonyl oxygen and an amide hydrogen in the peptide backbone would correspond to a volume increase of about 2.4 cm³/mol () compared with a non-hydrogen-bonded structure. Then, 2.2 cm³/mol () are regained upon formation of backbone–water or nonspecific intramolecular hydrogen bonds in the unfolded state (Fig. 5). We cannot exclude the possibility that steric effects and unfavorable packing contribute to the large transition state volume, but this seems unlikely to be a major factor, because only a single amino acid at the helix–coil border changes its ϕ,ψ angles and there is an energetically favorable and sterically allowed route through ϕ,ψ space from the unfolded–extended region to the helical region (37).

The volume decrease upon helix formation observed in TTET experiments is in agreement with FTIR studies on Ala-based peptides, which also observed a stabilization of the helical state (17). Using an equilibrium linear Ising model to analyze the FTIR data yielded a reaction volume of −0.98 ± 0.04 cm³/mol per residue, which is larger than the reaction volume of −0.23 cm³/mol per residue determined by TTET (Fig. 3). This discrepancy may be due to the difference in temperature between the FTIR studies (25.4 °C) and the TTET experiments (5 °C), because it is commonly observed that reaction volumes for protein folding are strongly temperature dependent (2, 3). The volume decrease upon formation of intramolecular helical hydrogen bonds is further in agreement with results on collagen folding, which showed a volume decrease upon triple helix formation (38). As in α-helical structures, formation of the collagen triple helix mainly leads to formation of solvent-accessible hydrogen bonds between carbonyl oxygens and amide protons in the peptide backbone.

Comparison of the reaction and activation volumes for adding–removing a single helical residue, , , and , with the respective volume changes for folding and unfolding of the helix in the center between the labels, , , and , shows that the experimentally determined values are apparent reaction and activation volumes that do not represent properties of individual kinetic steps during growth and shrinking of a helix. Due to the non–two-state nature of the helix–coil transition, the measured folding and unfolding rate constants are influenced by many microscopic steps of adding or removing individual helical residues (20, 22, 32, 33). The Monte Carlo simulations reveal that this mechanism leads to an increase in the reduced folding rate constant (k_f/k₁) and a decrease in the reduced unfolding rate constant (k_u/k₁) in the helix center with increasing pressure, i.e., with increasing s-value (Fig. 4). As a result, scaling of the pressure dependence of the reduced rate constants to the experimentally observed values gives = 2.2 cm³/mol compared with of 1.7 cm³/mol. At the same time, scaling leads to a decrease from = 3.3 cm³/mol to = 2.4 cm³/mol (Fig. 4). These opposing effects of the scaling factors give rise to the observed large difference in reaction volumes for helix stability in the center ( = −1.6 cm³/mol) and for adding a single helical residue ( = −0.23 cm³/mol). These comparisons reveal that kinetic coupling between the microscopic shrinking and growth steps at the individual positions in the peptide leads to macroscopically observable rate constants for helix folding and unfolding that cannot be analyzed by a simple two-state model.

Effect of Pressure on Loop Formation in the Unfolded State.

The analysis of the TTET experiments shows that loop formation in the unfolded or partially unfolded state of the helical peptides (k_c) is pressure dependent with an activation volume of 3.6 cm³/mol (Fig. 2). To test whether this large activation volume is characteristic for loop formation in an unfolded Ala-based polypeptide chain, we investigated the pressure dependence of the dynamics of loop formation in an 11-amino-acid peptide that corresponds to the central 11 amino acids in the 21-amino-acid helical peptide including the TTET labels

Leaving both the N- and the C terminus unprotected prevents formation of detectable amounts of helical structure as judged by its CD spectrum (Fig. S4A). Further, the temperature dependence of the CD signal at 222 nm reveals a monotonous decrease in ellipticity with increasing temperature (Fig. S4B), which is characteristic for an ensemble of unfolded conformations (39–41). A positive CD band around 227 nm and a negative band around 198 nm indicate the presence of significant amounts of polyproline II (PPII) structure in the ensemble of unfolded states of the 11-amino-acid peptide (42, 43), in agreement with the observation of PPII structure in a seven-alanine residue peptide (44).

Loop formation in the 11-amino-acid unfolded polypeptide exhibits only a weak pressure dependence (Fig. S4C) with an activation volume of 0.51 ± 0.09 cm³/mol, determined in a global fit of the kinetic traces (Fig. S4D). This result shows that the underlying fast conformational transitions in the unfolded state of an Ala-based peptide do not involve major volume changes. Comparison of these results with the pressure dependence of loop formation reaction (k_c) in the helical peptide shows that the large activation volume for loop formation of the unfolded state (U) is not characteristic for an unfolded polypeptide chain. The results rather suggest that only partial local unfolding of the helical structure between the labels is required to enable donor and acceptor to come into van der Waals contact, as assumed in our previous work (20, 22). Loop formation in the partially unfolded coil state of the peptide requires structural rearrangements that are associated with volume changes.

Conclusions

The TTET experiments on the helix–coil dynamics in helical peptides reveal that the volume decreases by 0.23 cm³/mol (0.38 ų per molecule) per helical residue upon helix formation. This volume change for formation of a single helical residue is small, but the coupling of many of these steps leads to a significant stabilization of the central part of a 21-amino-acid helix by pressure with a reaction volume for unfolding of −1.6 cm³/mol (Fig. 3A). Thus, the commonly observed increase in volume upon formation of the native state is not due to the formation of helical secondary structure, consistent with the idea that the volume increase upon protein folding is mainly due to void volumes formed in the native state (8). Notably, the actual void volumes in folded proteins are larger than the observed volume increase upon folding, because helix formation decreases the protein volume. Our results further reveal that the decrease in helix stability in folded proteins with increasing pressure, as observed in hydrogen–deuterium exchange studies (45), is not due to an effect of pressure on intrinsic helix stability. Rather, increasing pressure destabilizes the tertiary structure of proteins, which, in turn, destabilizes helical structures due to a cooperative destabilization of the folded state. Equilibrium intermediates with residual helical structure have frequently been observed in pressure-induced unfolding experiments monitored by NMR (19, 46–48). The population of these persisting helical structure can be explained by the opposing effects of pressure on the stability of tertiary and helical structure. Whereas the native structure becomes destabilized with increasing pressure, helices become more stable. Thus, sequences with high intrinsic helix propensity may remain helical even in the absence of long-range tertiary interactions and even become stabilized with increasing pressure. In addition, pressure may induce helix formation in sequences with marginal intrinsic helix stability. This effect is in contrast to thermal- and denaturant-induced equilibrium unfolding transitions of small proteins, which typically exhibit two-state behavior because both tertiary interactions and helical structures are destabilized under such conditions (30, 49, 50). Residual helical structure is also frequently observed in cold-denatured and alcohol-unfolded states of proteins. Similar to high pressure, these conditions lead to a destabilization of tertiary structure but to a stabilization of helices.

The results further reveal that the transition state for adding or removing a single helical residue has a 2.2 cm³/mol and 2.4 cm³/mol larger volume compared with either the helical or coil state, respectively. Thus, even small structural rearrangements in the polypeptide chain may proceed through high-energy states with large volumes, because large activation volumes are already observed for changing only a single pair of ϕ, ψ angles in a secondary structural element. These effects will contribute to the volume of protein folding transition states, which are typically large and often exceed the volume of the native state (9–12).

Materials and Methods

All peptides were synthesized and purified as described (20, 22). TTET measurements were performed as described (20, 22). The laserflash setup was modified for high-pressure measurements by introducing the High Pressure Cell System from ISS.

All measurements were performed in degassed 10 mM potassium phosphate buffer, pH 7 at 5 °C. Peptide concentrations were about 50 µM measured by xanthone absorbance at 343 nm (ε = 3,900 M⁻¹⋅cm⁻¹). Typically, six kinetic traces were averaged at each pressure. The microscopic rate constants k_f, k_u, and k_c were obtained from globally fitting all kinetic traces over the complete pressure range assuming a linear effect of the pressure on the logarithms of the microscopic rate constants (SI Text). The program ProFit (QuantumSoft) was used for data fitting.

Kinetic Monte Carlo simulations based on the linear Ising model were carried out using Eqs. 5–9 as described (20, 22) with the parameters given in the text. Simulations were performed using the framework of Matlab R2011 (MathWorks) in combination with compiled C code for time-consuming iterations. A detailed description of the simulation procedure is given in SI Text.