893

Journal of Biomolecular Structure & Dynamics, ISSN 0739-1102 Volume 29, Issue Number 5, (2012) ©Adenine Press (2012)

*Corresponding author:Prof. Ramakrishna V. HosurPhone (lab): 191-22-22782271 Phone (off): 191-22-22782488Fax: 191-22-22804610 E-mail: hosur@tifr.res.in

Aditi Borkar1#

Manoj Kumar Rout1

Ramakrishna V. Hosur1,2*

1Department of Chemical Sciences,

Tata Institute of Fundamental Research,

Homi Bhabha Road, Colaba,

Mumbai-400005, India2UM-DAE Centre for Excellence in Basic

Sciences, Mumbai University Campus,

Kalina, Santa Cruz Mumbai-400098,

India#Current affiliation – Department of

Chemistry, University of Cambridge,

Lensfield Road, Cambridge CB2 1EW,

UK

Denaturation of HIV-1 Protease (PR) Monomer by Acetic Acid: Mechanistic and Trajectory Insights from Molecular Dynamics Simulations and NMR

http://www.jbsdonline.com

Abstract

Inside a living cell there can be a variety of interactions for any given protein, which serve to regulate denaturation and renaturation processes. Insights into some of them can be obtained by in vitro studies using various denaturing agents. In this study, all-atom MD simulations in explicit solvent and NMR relaxation studies were performed on HIV-1 Protease (PR) in 9 M acetic acid (AcOH) (the commonly used denaturant during PR preparation). Following previous reports that denaturation proceeds via dissociation of the dimer into monomers, unfolding of the monomer by acetic acid has been explicitly investigated here. Direct visu-alization of the denaturation process and evidence for the mechanism of denaturation have been presented. Our simulations reveal that the denaturation of the PR monomer is caused due to direct interaction between acetic acid molecules and PR. Autocorrelation of N-H vectors calculated from the simulations have revealed that the α-helix and the surrounding β-strands represent the sensitive regions of the PR that respond maximally to the change in the solvent environment around the PR and are prone to disruption by acetic acid. This disruption is caused due to increased penetration of the acetic acid molecules into the PR structure by formation of preferred tertiary contacts and hydrogen bonds between the PR and acetic acid molecules. Following the loss of these critical interactions, the PR follows a random and non-equilibrating path on the conformation landscape and cycles between different denatured extended and compact states.

Introduction

Reversible denaturation-renaturation plays a very important regulatory role in a protein’s function. This is clearly dependent on the nature of the denaturant (1), its influence on the degree of denaturation and pathways of denaturation and rena-turation. This in turn is dependent on specific interactions between the protein and the denaturant under question (2-4). Inside a living cell there can be a variety of interactions for any given protein, which serve to regulate denaturation and renatur-ation processes. While it is difficult to study these interactions explicitly at residue level detail in vivo with the current level of technological developments, insights into some of them can be obtained by in vitro studies using various chemical and physical denaturing agents. Nuclear Magnetic Resonance (NMR) spectroscopy has been particularly successful in these efforts (5-7). However, in most cases NMR has provided information on equilibrium situations (8-14) and do not necessarily provide insights into the trajectories of these denaturation and renaturation pro-cesses. In this context Molecular Dynamics (MD) simulation in explicit solvent is an extremely powerful tool (15-22). It not only helps in unraveling the trajectories, but can also lead to insights into equilibrium states, which can be compared with NMR experimental results.

Open Access ArticleThe authors, the publisher, and the right holders grant the right to use, reproduce, and disseminate the work in digital form to all users.

894

Borkar et al.

Many molecular dynamics studies on PR mechanism (23-25), drug resistance (26-28) and unfolding (29-33) have previously been reported. Here, while the mech-anism and drug resistance simulations generally employ liganded PR as the starting structure, the unfolding simulations were performed with higher temperature as the denaturing condition. Extensive all-atom 100 ns simulation of both the monomeric and dimeric PR has shown greater flexibility of the termini in the monomer (29)

and decoupling between monomer folding and dimerization (30). Simulation of the 99-residue monomer in water at different temperatures revealed early assembly of the N and C termini into stable β-sheet structures (31). Apart from the termini, the ‘flaps’ too have been shown to be highly flexible regions of the PR (32, 33). The flap dynamics is considered to be essential in target entry and exit from the PR active site and corresponding open and closed conformations are observed in the crystal structures. MD simulation has also revealed that the unliganded protease predominantly populates the semi open conformation, with closed and fully open structures being a minor component of the overall ensemble (32).

In this study, we report all-atom-MD and NMR investigations on denaturation of HIV-PR by acetic acid. Acetic acid is a commonly used denaturant in in vitro stud-ies. When PR is expressed in E. coli cells, the protease comes in the inclusion bodies. After extraction by cell lysis, the PR in inclusion bodies is dissolved in strong acetic acid (strong denaturing condition) and refolded by dilution (34). Acetic acid interaction with the PR monomer thus would be important to gain understanding of denaturation and folding mechanisms of the protein.

One effect of acetic acid on the protein is to protonate its exposed acidic residues. Other effects would be to alter the conformation and stability of the protein. The pro-tonation chemical equilibrium of the protein would be described by the equation:

B2 1 AH 5 BH 1 A2 ... [1]

Where, B2 5 unprotonated PR, AH 5 Acetic acid molecule, BH 5 Protonated PR, A2 5 Acetate ion. The extent of dissociation of the acid would depend upon the concentration of the acid and consequently the pH of the solution.

Regardless of the position of the equilibrium, it is clear that in this system all the four species are free to interact in a solution and it is impossible to study any one of these interactions explicitly by experimental methods like NMR. However, hypothetical models can be generated in silico and since current MD simulation protocols cannot account for proton exchange, study of such isolated interactions becomes possible by molecular dynamics simulation in explicit solvent.

In an earlier investigation, we have described MD studies on the interaction of AH with BH. In these studies, the mature PR dimer and monomer were both protonated and simulated separately in 9M AcOH to visualize the denaturation of PR in acetic acid solution. Even at such high concentration of the acid (pH 1.9), the concentra-tion of acetate ions would be extremely small (pKa of acetic acid is 4.76) and thus the major interactions would be between the undissociated acid and the protein only. It was observed in a previous study by us (35) that the PR denaturation begins by sepa-ration of the dimer into intact monomers and then the monomers start to denature. However, the length of the simulation (30 ns for the monomer) was not long enough to see complete dissociation and hence complete denaturation of the protein. Con-tinuing these efforts, we have chosen to investigate here the monomer denaturation explicitly. For this we have considered the left hand side (L. H. S.) of the equilib-rium 1 above, namely, the interaction of unprotonated PR monomer with acetic acid. Apart from visualizing PR monomer denaturation, this will also help to unravel if protonation has any role to play in specific interaction with acetic acid and thus in the mechanism of denaturation. The insights from the simulations have been compared with NMR relaxation measurements, which throw light on dynamics at nanosecond

895

Denaturation of HIV-1 Protease: MD and NMR

Studies

time scale. Stepwise unfolding leads to differential changes in the dynamics along the protein chain and this will show up in the NMR measurements.

Methods

Construction of 9 M Acetic Acid Model

Both uncharged and charged acetic acid molecules were constructed using Argus-Lab software [http://www.arguslab.com] and their parameters were obtained from PRODRG server at Dundee (36) [http://davapc1.bioch.dundee.ac.uk/prodrg/]. The uncharged molecule was duplicated and a box [5 Å 3 5 Å 3 5 Å] containing 334 molecules of acetic acid was constructed using Vega ZZ 2.3.2 (37) [http://www.ddl.unimi.it]. All the further steps in setting up of the model were then carried out by GROMACS 4.0.5 (38, 39) tool and all MD simulations were performed in peri-odic boundary conditions (PBC) using GROMACS 4.0.5 tool with GROMOS96 force field (40). The acetic acid box was solvated with SPC216 water molecules

(41). The ratio of number of acetic acid molecules to that of water in the solvated box was adjusted to 1:3 and the density was adjusted to 1.02 kg l21 such that this box models 9 M AcOH. The Ka for acetic acid is 1.74 3 1025. Hence at pH 1.9 in pure water, the ratio of number of acetate ions to that of acetic acid molecules is 1.4:1000. Thus, in the 9 M AcOH model constructed above, all acetic acid mol-ecules were kept in protonated and uncharged form. The system composition is out-lined in Table I. Since the force field parameters for acetic acid molecules obtained in the GROMACS topology from PRODRG server and those for the SPC216 water molecules are already validated, we used these same validated force field param-eters for the simulation of the 9 M AcOH solvent.

After an initial energy minimization by steepest descent algorithm with tolerance of 100 kJ mol21 nm21 to remove any bad contacts, an MD run was set up for 1 ns to completely mix and equilibrate the contents of the box. Complete mixing and equilibration of the contents was confirmed by convergence of potential energy of the system and by convergence of the number of hydrogen bonds between acetic acid and water molecules using GROMACS 4.0.5 analysis tools.

Model Preparation and MD Setup for Studying Denaturation of PR

All the steps in setting up of the model for simulation were carried out by GROMACS 4.0.5 tool and all MD simulations were performed in PBC using GROMACS 4.0.5 tool with GROMOS96 force field. The solution structure of PR monomer reported by Ishima et al., (2003) (42) (PDB ID 1Q9P) was adopted as the starting struc-ture. This was placed in a cubic box and solvated separately with the equilibrated 9 M AcOH solution and SPC216 water. An initial energy minimization by steep-est descent algorithm with tolerance of 100 kJ mol21 nm21 was performed on this system to remove any bad contacts. This was followed by 50 ps of MD with posi-tion restraints on the PR to relax the water and AcOH in the system. A 150 ns MD without any restraints was then performed on the system of monomeric PR in 9 M AcOH box and a 100 ns MD without any restraints was performed on the system of monomeric PR in water using 2 fs integration time and a cutoff of 1.2 nm for long-range interactions. Temperature was coupled to an external bath of 300 K. All trajectory analyses were carried out by the analysis tools in GROMACS 4.0.5 package and VMD (43) [http://www.ks.uiuc.edu/Research/vmd/] was used for visualization purposes.

NMR 15N Longitudinal Relaxation Measurements

NMR experiments were performed at 25ºC on Bruker Avance 800 MHz spectrometer. Longitudinal relaxation rates (R1) were measured using the pulse sequences described by Farrow et al., (1994) (44) with the following

Table IComposition of the system simulated in this study.

S. No.Name of the chemical species

Quantity in the simulation model

1 Unprotonated PR molecule

Single

2 Acetic Acid Molecules 7023 Water Molecules 21104 Chloride ions 3

896

Borkar et al.

inversion recovery delays: 10, 30, 60, 100, 200, 300, 450, 600, 800 and 1000 ms. At the end of these, an HSQC spectrum was again recorded to check for the stabil-ity of the protein. We observed no change in the HSQC spectrum indicating that the protein was very stable and also had reached equilibrium state at the beginning of the experiments. For all the two dimensional relaxation experiments, 2048 and 256 complex points were used along the t2 and t1 dimensions, respectively. All the data were processed using CARA and FELIX. The R1 values were extracted by fitting the peak intensities to the equation, I(t) 5 B exp (2R1 t).

Results and Discussion

In Acetic Acid Solution, PR Undergoes Renaturation-Denaturation Transitions

For a denatured heteropolymer, the radius of gyration RG can be estimated from the number of residues (45-47) and the RG for denatured PR is calculated to be around 3.1 nm. The RG value for the unprotonated PR in 9 M AcOH MD simulation [Figure 1(B)] approach this random coil behavior around 80 ns and after further 10 ns again fall off towards a more compact structure. In contrast, the RG value for the PR in water does not fluctuate much and remains approximately at the initial value [Figure 1(E)]. The RMSD value [Figure 1(A)] too rises steeply within the 1st ns of the simulation and then keeps on fluctuating between 0.5 nm to 0.7 nm. Around 65 ns it dips up to 0.4 nm but rises again around 80 ns. On the other hand, the

Figure 1: RMSD and RG analysis of the PR simulation and (A, B, C) 9 M AcOH and (D, E, F) water. The RMSD of the backbone atoms in the trajectory of unprotonated PR in (A) 9 M AcOH and in (D) water from the backbone atoms of the NMR structure of PR does not equilibrate and fluctuates more than 10 Å indicating a non-equilibrium run. Further, the radii of gyration with respect to the X, Y, Z axes of unprotonated monomeric PR during the 150 ns simulation in (B) 9 M AcOH and in (E) water is shown in the Figure. Finally, the RMSD of the backbone atoms in the trajectory from the backbone atoms of the immediately previous structure in (E) 9 M AcOH and (F) water is also reported. In these Figures, solid black lines report the equilibrating run in water. Solid gray lines represent the path followed by the unpro-tonated PR molecule in 9 M AcOH. The Broken black lines are running average of the backbone RMSD over every 5 ns, which gives a general trend of the structural transitions of the protein during this simulation.

897

Denaturation of HIV-1 Protease: MD and NMR

Studies

backbone RMSD of the PR in water simulation equilibrates around 0.32 nm. Thus, it can be concluded that during this simulation, PR undergoes denaturation and then fluctuates between a state closer to the native state and more unfolded states. Such transitions were not seen in the comparatively shorter 30 ns run of protonated PR in 9 M AcOH, which were previously performed by us (35).

The plot of backbone RMSD to the immediately previous structure [Figure 1(C) and Figure 1(F)] also reinstates the fact that the PR structure does not equilibrate but changes continuously in the simulation conditions. Up to the first 10 ns, the RMSD values for PR simulations in 9 M AcOH and water are similar and the struc-ture of PR does not deviate much; but immediately afterwards, a major structural transition in the 9 M AcOH simulation causes the protein to enter a highly dynamic path. Hence, it is expected that at this point in the trajectory, some major stabiliz-ing interaction is lost in the protein structure. Contrary to this, no such drastic changes appear anywhere for the full simulation of PR in water. This plot also shows the general trend in the structural transitions of PR in 9 M AcOH. Each drastic structural transition is followed by a trough that tries to maintain the new interactions formed in the trajectory; however, such interactions remain transitory and again propel the structure into further search of a stabilizing conformation. It is not prudent to conclude that the troughs correspond to possible intermediates on the folding pathway. This can only be established after free energy calculations for this trajectory.

Residue Level Dynamics of PR in Acetic Acid Solution

The nature of the autocorrelation function, C(t), plotted for the N-H vectors for each residue of a protein gives an indication of the motion and hydrogen bonding flexibility of the residue in the protein. Rapidly decaying C(t) is representative of fast backbone dihedral angle fluctuations of the residue (48, 49) while slow con-tinuously decaying C(t) denotes that the time scale of motion of the N-H vector is comparative to the global motion of the protein (48). The pattern of the C(t) cal-culated for the residues in PR simulation in 9 M AcOH and water can be divided into 3 types: [1] decaying C(t) that converges [2] a continuously decaying C(t) that does not converge and [3] fluctuating C(t) that does not resemble any of the previous two profiles [supporting information Figure S1]. The asymptotic limit of the first type of C(t) gives the order parameter, S2, of the residue. The second type of profile indicates an N-H vector whose fluctuations are comparative to the motions of the region of the protein backbone on which it is sitting. Hence, it would also be expected to form part of a stable hydrogen bonded structure. The third profile is indicative of several types of internal motions exhibited by the N-H vector and thus represents a residue involved in conformational exchange

(48, 50). While the C(t) of the first two types do not convey much structural information, the residues exhibiting the C(t) of the third type could be considered as the hot-spots for the conformational dynamics of PR (50). Such results become more interesting when longer contiguous stretches of residues follow the same pattern instead of isolated residues.

Majority of the residues exhibit C(t) of the first type for simulation of PR in both water and 9 M AcOH [data not reported]. But there are some stretches of residues that completely alter their behavior in acetic acid as compared to that in water [Figure 2(A) and 2(B)]. For example, consider the α-helical stretch from 87-94 and the residues 30-32. In water [Figure 2(B)], most residues in both these stretches show C(t) of the second type and hence indicate a stable hydrogen bonded structural element of PR (i.e., α-helical for the first case). However, in 9 M AcOH [Figure 2(A)], the α-helical stretch exhibits C(t) of the third type indi-cating their importance in the structural dynamics of PR in acetic acid whereas residues 30-32 show C(t) of the first type indicating loss of the strong hydrogen bonded interactions with some particular region of the protein.

898

Borkar et al.

In other words, the stretches outlined above comprise of the sensitive regions of the PR. They respond significantly to the change in solvent environment of the PR and contribute to the different structural dynamics of the PR seen in acetic acid as compared to water. Interestingly, the spatially closed unit of α-helix (residues 83-92) with sheet (residues 74-78) above and another β-strand with sheet (residues 24-34) perpendicular to these elements form the critical folding nucleus of the PR

(51, 52). Hence, the sensitivity and loss of stable H-bonding of the α-helical and 30-32 residues along with small time scale C(t) observed for residues 74-78 make this region of the PR structurally unstable and an important target for denaturation by acetic acid.

Longitudinal 15N Relaxation Rate Studies on PR

Longitudinal 15N relaxation rate (R1) reports on fluctuations and motions along the chain occurring at nanosecond time scale. Higher rates indicate greater flexibility and lower rates represent greater rigidity of the protein backbone. At 9 M AcOH, the PR exists as a monomer in solution for NMR experimentation. Figure 3(A) enlists the residue-wise R1 values measured for PR in 9 M AcOH and Figure 3(B) shows these in a color-coded manner on the PR monomer structure. Clearly, the R1 values for the α-helical residues and β-strand 8 just before the helix (represented as

Figure 2: Representative autocorrelation function C(t) for N-H vectors during PR simulation in (A) 9 M AcOH and (B) water for three residues Ile 64, Leu 90 and Asp 30. These residues form parts of the PR stretches that are predicted to be sensitive to acetic acid environment around the PR. These residues correspondingly show completely altered behavior in acetic acid and water. For details, refer to the text.

Figure 3: (A) Residue wise longitudinal relaxation rates (R1) of PR in 9 M AcOH. A1, A2 and A3 are regions having R1 values higher than the average. These values are shown in blue color on the structure of the native protein in (B). The regions B1 and B2 (red) are the stretches having R1 values lower than the average.

899

Denaturation of HIV-1 Protease: MD and NMR

Studies

region A3 in Figure 3(A)) are higher than average for the whole structure indicating that the helix is a flexible and a vulnerable region of the PR. Likewise, the region 50-70 (B2 in the Figure) which is involved in a long β-sheet on the outer surface of the protein in the native structure displays conspicuously low R1 values implying greater rigidity in this stretch.

Insights into the Structural Changes Accompanying Denaturation of PR

To probe into this matter further, secondary structure profiles [Figure 4] of this trajectory were studied. These calculations show that indeed the C-terminal α-helix transits to a turn within 7 ns of the simulation and subsequently, a β-sheet appears in that region around 28 ns. Yan and co-workers (2008) (31) too have reported the assembly of the N- and C-termini of the PR monomer into a β-sheet during a high temperature MD simulation. However, in our studies, the N-terminal residues remain as random coils.

Tertiary contact profiles (a tertiary contact is defined between two residues when the non-bonded heavy atoms of the residues lie within 0.6 nm of each other) of PR in 9 M AcOH [Figure 5(A) and 5(B)] at 10 ns also reveal that the α-helix interac-tions vanish and these residues instead show enhanced contacts with the proximal residues of β-strands 5 and 6 but not with the loop connecting these two strands. Although the new contacts appear in the water simulation, [supporting informa-tion Figure S2], there is no loss of α-helical interactions throughout the trajectory. Thus, from all these observations, one concludes that the distinguishing and major structural transition that initiates acid denaturation of PR monomer is the loss of the α-helix. This is supported by the NMR observation in 9 M AcOH that the region corresponding to the helix in the native protein structure is the most flexible region along the chain, which indicates that the structure around this has been most dis-turbed by acetic acid interaction.

Evidence of Direct Interactions of Acetic Acid with PR

Figure 6 illustrates the radial distribution function of both acetic acid and water around the PR. It is clearly seen that there is more probability of finding the

Figure 4: The secondary structure profile calculated using DSSP (53) for the initial 30 ns of the two simulations of monomeric PR in (A) unprotonated state in 9 M AcOH and in (B) water (35).

900

Borkar et al.

acetic acid molecules around the PR than water. Also, there is no long-range order between the acetic acid molecules. Hence they can be seen to be preferentially residing closer to the PR in the box during the simulation.

Secondly, if the PR forms unbiased contacts with the solvent molecules and does not prefer one chemical species over the other, then probability of PR contacting water molecules will be more than that of PR contacting acetic acid molecules in 9 M AcOH because of the smaller size and larger number of water molecules in the solution than acetic acid molecules (the ratio of number of molecules of water to that of acetic acid is 3:1). However, Figures 7(A) and 7(B) tell a completely dif-ferent story. The number of tertiary contacts between PR and water drops sharply within the first few nanoseconds of the simulation and then equilibrates. On the other hand, the number of tertiary contacts between PR and acetic acid immediately ascend (within the same period) to and then equilibrate at a value that is twice that of the equilibrium value in water. The same can be said for the hydrogen bonding preferences of PR [Figures 7(C) and 7(D)]. In the presence of acetic acid, the protein forms one and a half times more number of hydrogen bonds with acetic acid as compared to water. Hence, indeed the PR forms preferred contacts with the acetic acid molecules in 9 M AcOH.

Figure 8 shows the number of tertiary contacts formed by only the helical resi-dues of PR with acetic acid and water. Clearly, the tertiary contacts and hence the interactions between these residues and acetic acid increase steeply within 10 ns of

Figure 5: Tertiary contacts at (A) 0 ns and (B) 10 ns and snapshot structures at the same time points i.e., at (C) 0 ns and (D) 10 ns of the simulation of the unprotonated monomeric PR in 9 M AcOH. No α-helical interactions are seen in the tertiary contact profiles as pointed out by the arrows in the Figure. Similarly, the snapshot structure at 10 ns shows a random coil at the residues that form a good α-helix in the native PR. For a detailed description of the tertiary contact profiles, refer to supporting informa-tion Figure S2.

Figure 6: The radial distribution function of acetic acid and water around the PR during the simulation. Acetic acid is clearly more concentrated around the PR than water. It also, unlike water, shows no long-range order.

901

Denaturation of HIV-1 Protease: MD and NMR

Studies

the simulation whereas those with water do not change significantly and keep on fluctuating. Interestingly, the loss of the α-helical structure as seen from Figure 4 and the loss of critical intra-helix and helix-β-sheet interactions in PR as seen from Figures 5(A) and 5(B) are concomitant with this first surge in the number of tertiary contacts of helical residues with acetic acid. Thus, the disruption of these interac-tions by acetic acid causes the initiation of denaturation of the PR. The second surge in the number of tertiary contacts of helical residues with acetic acid is concomitant with the formation of β-sheet near the C-terminal of PR in acetic acid simulation.

All these observations on unprotonated PR in 9 M AcOH point out the fact that it is the direct interaction of acetic acid molecules with PR that causes the PR to denature. Similar results were observed in the case of protonated PR as well (35). In either case denaturation is initiated with the vanishing of the α-helical interactions and it is the interaction of the PR with AcOH that drives the unfolding. Only in case of unprotonated conditions, the PR seems to be more unstable and starts denaturing within 7 ns and also forms β-sheet at the C-terminal.

Conclusions

We have presented here the first description of the mechanism of denaturation of PR by acetic acid using all atom MD simulation in explicit solvent and NMR relax-ation studies. Autocorrelation of N-H vectors and NMR longitudinal 15N relaxation measurements have revealed that the α-helix and the surrounding β-strands rep-resent the sensitive regions of the PR that respond maximally to the change in the solvent environment around the PR. The PR monomer denaturation in 9 M AcOH indeed then is shown to be initiated by disruption of the folding nucleus formed by these stretches due to increased penetration of the acetic acid molecules in to the PR structure. The PR in turn allows for this penetration due to the formation of preferred tertiary contacts and hydrogen bonds between the PR and acetic acid molecules. However, during the time duration monitored in this study, only some early events could be observed. Nevertheless, these provide satisfactory evidence

Figure 7: The total number of tertiary contacts made by the complete PR with (A) water and (B) acetic acid and the total number of hydrogen bonds formed by the complete PR with (C) water and (D) acetic acid during the 150 ns simulation of monomer PR in 9 M AcOH.

Figure 8: The total number of hydrogen bonds formed by the α-helical residues 85-95 of the PR with water and acetic acid in 9 M AcOH.

902

Borkar et al.

of the mechanism of denaturation. Simulations up to microseconds may be required to visualize collapse of the robust β-sheet structure and complete denaturation of the protein. Such studies can also be extended to denaturation by other agents to obtain insights into the overall denaturation landscape of the protein.

Supplementary Material

Supplementary material dealing with Representative autocorrelation function C(t) for N-H vectors during PR simulation and the tertiary contacts between the residues of monomeric PR simulated in water is available at no charge from the authors directly; the supplementary data can also be purchased from Adenine Press for US $50.00.

References

C. Camilloni, 1. et al., Biophysical Journal 94(12), 4654-4661 (2008).M. C. Stumpe and H. Grubmüller. 2. PLoS Comp Biol 4(11), e1000221 (2008).A. Das and C. Mukhopadhya. 3. J Phys Chem B 112(26), 7903-7908 (2008).Z. Zhang, Y. Zhu, and Y. Shi. 4. Biophys Chem 89(2-3), 145-162 (2001).P. Schanda, V. Forge, and V. Brutscher. 5. Proc Natl Acad Sci USA 104(27), 11257-11262 (2001).M. Zeeb and J. Balbach. 6. Methods 34, 65-74 (2004).N. A. J. Van Nuland, V. Forge, J. Balbach, and C. M. Dobson. 7. Acc Chem Res 31, 773-780 (1998).A. G. Palmer. 8. Chem Rev 104, 3623-3640 (2004).A. Mittermaier and L. E. Kay. 9. Science 312, 224-228 (2006).K. Schlepckow, J. Wirner, A. Bachmann, T. Kiefhaber, H. Schwalbe. 10. J Mol Biol 378, 686-698 (2008).L. Hoang, S. Bédard, M. M. G. Krishna, Y. Lin, and S. W. Englander. 11. Proc Natl Acad Sci USA 99, 12173-12178 (2002).G. S. Huang and T. G. Oas. 12. Proc Natl Acad Sci USA 92, 6878-6882 (1995).M. Wang, Y. Tang, S. Sato, L. Vugmeyster, C. J. McKnight, and D. P. Raleigh. 13. J Am Chem Soc 125, 6032-6033 (2003).D. M. Korzhnev, X. Salvatella, M. Vendruscolo, A. A. Di Nardo, A. R. Davidson, 14. C. M. Dobson, and L. E. Kay. Nature 430, 586-590 (2004).P. L. Freddolino, F. Liu, M. Gruebele, and K. Schulten. 15. Biophysical Journal 94(10), L75-L77 (2008). Y. Duan and P. Kollman. 16. Science 282, 740-744 (1998).R. D. Schaeffer, A. Fersht, and V. Daggett. 17. Curr Opin Struc Biol 18, 4-9 (2008).V. S. Pande, I. Baker, J. Chapman, S. P. Elmer, S. Khaliq, S. M. Larson, Y. M. Rhee, 18. M. R. Shirts, C. D. Snow, and E. J. Sorin, et al., Biopolymers 68, 91-109 (2003).W. Zhang, C. Wu, and Y. Duan. 19. J Chem Phys 123, 154105 (2005).G. Settanni, J. Gsponer, and A. Caflisch. 20. Biophyisical Journal 86, 1691-1701 (2004).E. Paci, K. Lindorff-Larsen, C. M. Dobson, M. Karplus, and M. Vendruscolo. 21. J Mol Biol 352, 495-500 (2005).K. Lindorff-Larsen, M. Vendruscolo, E. Paci, and C. M. Dobson. 22. Nat Struct Mol Biol 11, 443-449 (2004).R. W. Harrison and I. T. Weber. 23. Protein Eng 7(11), 1353-1363 (1994).J. Trylska, V. Tozzini, C. Chang, and J. A. McCammon. 24. Biophysical Journal 92(12), 4179-4187, (2007).Z. Zhu, D. I. Schuster, and M. E. Tuckerman. 25. Biochem 42(5), 1326-1333 (2003).X. Chen, I. T. Weber, and R. W. Harrison. 26. J Mol Model 10(5), 373-381 (2004).F. Wartha, A. H. C. Horn, H. Meiselbach, and H. Sticht. 27. J Chem Theory Comput 1(2), 315-324 (2005).A. L. Perryman, J. H. Lin, and J. A. McCammon. 28. Protein Sci 13(4), 1108-1123 (2004).Y. Levy and A. Caflisch. 29. J Phys Chem B 107(13), 3068-3079 (2003).Y. Levy, A. Caflisch, J. N. Onuchic, and P. G. Wolynes. 30. J Mol Biol 340(1), 67-79 (2004).M. C. Yan,31. et al., Protein Struc Func Bioinfo 70(3), 731-738 (2008).V. Hornak, A. Okur, R. C. Rizzo, and C. Simmerling. 32. Proc Natl Acad Sci USA 103(4), 915-920 (2006).J. R. Collins, S. K. Burt, and J. W. Erickson. 33. Nat Struct Mol Biol 2(4), 334-338 (1995).J. O. Hui,34. et al., J Protein Chem 12(3), 323-327 (1993).A. Borkar, M. K. Rout, and R. V. Hosur. 35. PLoS One 6(6), e19830 (2011).A. W. Schuttelkopf and D. M. F. van Aalten. PRODRG: a tool for high-throughput 36. crystallography of protein-ligand complexes. Acta Crystallogr Sect D 60(8), 1355-1363 (2004).

903

Denaturation of HIV-1 Protease: MD and NMR

Studies

A. Pedretti, L. Villa, and G. Vistoli (2004). 37. J Comput-Aided Mol Des 18(3), 167-173 (2004).H. J. C. Berendsen, D. van der Spoel, and R. van Drunen. 38. Comput Phys Comm 91(1-3), 43-56 (1995).E. Lindahl, B. Hess, and D. van der Spoel. 39. J Mol Model 7(8), 306-317 (2001).L. D. Schuler, X. Daura, and W. F. van Gunsteren. 40. J Comput Chem 22(11), 1205-1218 (2001).H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, and J. Hermans.41. Interaction models for water in relation to protein hydration in intermolecular forces. Ed. Pullman, B. (Dordrecht The Netherlands: Reidel), 331-342 (1981).R. Ishima, D. A. Torchia, S. M. Lynch, A. M. Gronenborn, and J. M. Louis. 42. J Biol Chem 278(44), 43311-43319 (2003).W. Humphrey, A. Dalke, and K. Schulten. 43. J Mol Graphics 14(1), 33-38 (1996).N. A. Farrow,44. et al., Biochem 33(19), 5984-6003 (1994).I. S. Millett, S. Doniach, and K. W. Plaxco. 45. Adv Protein Chem 62, 241-262 (2002).P. J. Flory. Statistical Mechanics of Chain Molecules. Wiley, New York (1969).46. J. E. Kohn, 47. et al., Proc Natl Acad Sci USA 101(34), 12491-12496 (2004).K. V. Pervushin, V. Y. Orekhov, D. M. Korzhnev, and A. S. Arseniev. 48. J Biomol NMR 5(4), 383-396 (1995).S. Swaminathan, T. Ichiye, W. F. van Gunsteren, and M. Karplus. 49. Biochem 21(21), 5230-5241 (1982).I. Chandrasekhar, M. G. Clore, A. Szabo, A. M. Gronenborn, and B. R. Brooks. 50. J Mol Biol 226(1), 239-250 (1992).R. A. Broglia, G. Tiana, L. Sutto, D. Provasi, and F. Simona. 51. Protein Sci 14(10), 2668-2681 (2005).A. Wallqvist, G. W. Smythers, and D. G. Covell. 52. Protein Eng 11(11), 999-1005 (1998).W. Kabsch and C. Sander. 53. Biopolymers 22(12), 2577-2637 (1983).

Date Received: October 24, 2011

Communicated by the Editor Ramaswamy H. Sarma