A 12 layer water soak was then built around all atoms of the inhibitor, the metal ions, the terminal DNA dinucleotide, and IN residues near the active site . This solvated structure was Daunorubicin energy minimized using the AMBER99 force field with the distance dependent dielectric electrostatics function. Each simulation began at, and equilibrated to, 310 K for 100 ps with a time step of 0.002 ps. The NPT ensemble and NPA algorithms were employed . A total of 8 sets of simulations were run in duplicate for each calculation . Internal and interaction energies were separately calculated for the following components of the inhibitorbound complex: the IN residues listed above, DNA residues C16 and A17, the Mg2 ions, and the inhibitor. Energy calculations for each of these four atoms sets were run in duplicate, requiring 8 total simulations.
The same calculations were performed for the IN residues, the dinucleotide, andMg2 ions in the uninhibited complex. In this uninhibited model, the terminal adenosine that is displaced by INSTI binding is in the conformation required for strand transfer with its 3= OH bound to the Mg2 ion SGLT . Using just the coordinates of the inhibitor from the RAL IN DNA complex, a 10 layer water shell was generated around the ligand and energy minimized, as described for the complex, using the MMFF94x force field. In the INSTI bound intasome model, the 12 layer water soak covers all of the atoms luded in energy calculations. For the unbound ligand, a 10 layer water soak was sufficient to cover all the atoms in the calculation.
MD simulations for energy calculations of the protein and DNA were run using nonpositivist AMBER99, while simulations for energy calculations of the ligand and Mg2 ions were run using MMFF94x. The length of all simulations was 100 ps with a time step of 0.002 ps. Initial and final temperatures were 310 K. All water molecules from the soaks described above were kept explicit, and their bonds were held rigid. Internal and interaction potentials of each set of atoms were recorded every 0.5 ps. The first 50.0 ps was treated as an equilibration period and not luded in calculations, while the mean and standard deviation of both the internal and interaction potentials were calculated for the remaining 100 time points . The differences between the uninhibited and inhibited complexes gave the energy change of each component of the complex during the transition from the unbound INSTI to bound INSTI states.
Each simulation gave 100 values each for internal energy, interaction energy, and total energy , and mean values were calculated for each simulation. Each value reported here is the mean of the means of the duplicate simulations. Determining Mg2 interaction potentials. MD simulations were performed using the MMFF94x force field parameters, NPT ensemble, andNPAalgorithms with a starting and equilibrated temperature of 310K . As a reference, the two magnesium ions from our WT RAL model were stripped of surrounding protein, DNA, and inhibitor and solvated as described above. The interaction potential for these solvated Mg2 ions was determined using this MD approach. Hydration of Mg2 ions has been studied using more exhaustive molecular dynamics approaches, and the solvation energies have been reported to be in the range.