Dataset showing the impact of the protonation states on molecular dynamics of HIV protease

The data described here supports the research article “Unraveling HIV Protease Flaps Dynamics by Constant pH Molecular Dynamics Simulations” (Soares et al., 2016) [1]. The data involves both standard Molecular Dynamics (MD) and Constant pH Molecular Dynamics (CpHMD) to elucidate the effect of protonation states of catalytic dyad on the HIV-PR conformation. The data obtained from MD simulation demonstrate that the protonation state of the two aspartic acids (Asp25/Asp25′) has a strong influence on the dynamics of the HIV-PR. Regarding the CpHMD simulation, we performed pka calculations for HIV-PR and the data indicate that only one catalytic aspartate should be protonated.


a b s t r a c t
The data described here supports the research article "Unraveling HIV Protease Flaps Dynamics by Constant pH Molecular Dynamics Simulations" (Soares et al., 2016) [1]. The data involves both standard Molecular Dynamics (MD) and Constant pH Molecular Dynamics (CpHMD) to elucidate the effect of protonation states of catalytic dyad on the HIV-PR conformation. The data obtained from MD simulation demonstrate that the protonation state of the two aspartic acids (Asp25/Asp25 0 ) has a strong influence on the dynamics of the HIV-PR. Regarding the CpHMD simulation, we performed pk a calculations for HIV-PR and the data indicate that only one catalytic aspartate should be protonated. &

Data
The data presented here comprise results obtained from MD and CpHMD to investigate the functional role of protonation states on HIV-PR. The analyses that were carried out employed mainly three systems: free HIV-PR, HIV-PR complexed with a natural substrate (p1p6 peptide) and HIV-PR complexed with an inhibitor compound (Nelfinavir). We have measured the root mean square deviation profile in different protonation states using standard MD simulation ( Fig. 1), root mean square fluctuation in different pH values using CpHMD (Fig. 2, Table 1), protonation ratio of all ionizable residues (Fig. 3) and conformational cluster analysis (Fig. 4). Table 2 elucidates an unusually low predicted pK a for Glu35, through a hydrogen bond prevalence analysis.

Systems Setup
The atomic coordinate entries available in the PDB under accession codes 1OHR [6] and 1KJF [7] were used as initial condition of the NFV and p1/p6 bound-HIV-PR, respectively. The crystal structure 2HB4 was used as apo form of HIV-PR [8]. Before starting the simulation on the 1KJF structure, we performed the mutation N25D [1].

MD simulations
The nelfinavir-bound HIV-PR was simulated under three different protonation states in the catalytic residues: deprotonated (both residues are not protonated), monoprotonated (only the Asp25' is protonated) and deprotonated (both residues are protonated). We performed MD simulations with the program package GROMACS v. 4.4, using the GROMOS96 (43A1) force field and the SPC water model [9]. The solvation procedure was performed with a layer of at least 15 Å around the complex (approximately 11,350 water molecules). Due to their net positive charge, an appropriate number of chloride counter-ions were added to neutralize the system. The topology for nelfinavir was taken from a previous reference [10]. For Coulomb interactions, the reaction field correction term [11] was employed, with Cutoff 1.4 Å and a dielectric constant set to 54 [12]. All systems were run in periodic boundary conditions and the NPT ensemble. The temperature was maintained at 300 K and pressure at 1 atm using the Berendsen weak coupling approach [13]. LINCS and SETTLE were applied to constrain solute and solvent bonds respectively. Each initial set up was optimized using three different steps. First, an energy minimization using the steepest-descent algorithm was made restraining the protein and ligand heavy atoms to their original positions with a harmonic potential of 10 kcal mol À 1 Ǻ À 2 . Then, another minimization using steepest-descent with no restraints was performed. Finally, an energy minimization procedure with all restraints already removed was conducted, using the conjugate gradient method until reaching a gradient of 2.39 kcal mol À 1 Ǻ À 1 . Following the minimization, two stages of equilibration were performed: a 500 ps of MD with protein non-hydrogen atoms positions restrained by harmonic potential of 2 kcal mol À 1 Ǻ À 2 , in a NVT ensemble and other of 2 ns MD simulation with no position restraints, in a NPT ensemble. Finally, 60 ns of MD were conducted for further analysis. -(Dash) denotes that the pk a value was not possible to be calculated, because there was no enough change in the protonation of residue.

CpHMD simulations
The HIV-PR was simulated in its apo form, and bound to p1p6 and nelfinavir at pH 5.0 pH 6.0 and pH 7.0 for 40 ns and 80 ns. The CpHMD simulations were performed with the program package Amber12 and the AMBER ff99SB force field [5]. The parameters for nelfinavir were obtained using the antechamber module and the GAFF force field [14]. All systems were run using implicit solvation model (igb ¼2) [15][16][17] under NVT ensemble. Before the simulation only the Asp25' was protonated. For non-bonded interactions was employed a cutoff 30 Å. The system temperature was maintained at 300 K using the Berendsen thermostat [18]. All bond lengths involving hydrogen atoms were constrained using the SHAKE algorithm [19]. The same processes described in the previous section were also employed in this step, all Glu, Asp and His residues were allowed to change protonation states (18 residues). Protonation state change attempts were made every 10 fs. Fig. 4. Correlation between the conformational and protonation states for apo (panel A), p1p6-bound (panel B) and nelfinavirbound HIV-PR (panel C) at pH 7. The y axis represents the RMSD cluster number on the respective time frame (x axis). The blue, red and yellow bars represent deprotonated, protonation of Asp25 and protonation of Asp25', respectively. The Dashed lines point to the centroids of the three most prevalent clusters found throughout the simulation time of the apo system (clusters 1, 2 and 3). The clusters centroids of the remaining systems ("p1p6" and "nel") were not represented.