Accurate Prediction of Drug Activity by Computational Methods: Importance of Thermal Capacity

Abstract

Heat capacity is one of the most important thermodynamic quantities in protein biochemistry. Upon the binding of small molecules, a change in the heat capacity of proteins is generally observed, and this is often used in drug discovery. However, few computational works dedicated to the study of these phenomena are available in the literature. Here, a simple computational method for determining the change in heat capacity upon the binding of small ligands has been evaluated. The method is based on the accurate calibration of the solvent's thermal properties in the simulation conditions used in order to simply subtract its contribution to calculate the variations in the heat capacity of the system of interest. Using HIV protease as a model system, for which numerous experimental thermodynamic data are available, estimates of the change in heat capacity upon binding were obtained, which were similar to those observed experimentally. Furthermore, the predicted variations in heat capacity appear to be able to discriminate between molecules that behave as effective inhibitors of the enzyme and molecules that are able to bind the enzyme but not inhibit it. The results obtained suggest that this computational approach could be a useful aid in the in silico screening of new ligands for targets of interest.

Keywords: HIV; heath capacity; ligands; molecular dynamics simulation; protease; protein binding; thermodynamics; water.

Figure A1

Energy and temperature values obtained during a simulation. The data are those of one of the replicates of the HIV protease—indinavir complex at 310 K. (a) Temperature data. (b) Total energy data.

Figure A1

Energy and temperature values obtained during a simulation. The data are those of one of the replicates of the HIV protease—indinavir complex at 310 K. (a) Temperature data. (b) Total energy data.

Figure A2

Temperature distribution obtained during a simulation. The data (gray bars) are those of one of the replicates of the HIV protease—indinavir complex at 310 K; the Gaussian fit is reported as black solid line (mean value and standard deviation are reported at the top of the graph).

Figure A3

Redocking of the ligands. The crystallographic ligand is represented in lime, while the pose obtained by molecular docking is reported in orange. (a) MK1 and 1SDT. (b) 6EH and 5IVQ. (c) 6EF and 5IVS.

Figure 1

Calculated total energies for pure water. The data reported refer to a series of simulations (three replicates) for a water box containing 9138 water molecules. The line that best fits the data is also shown (the equation has the form $U_{tot}=169.1957T-123927.2490$, $R^2=0.9999$).

Figure 2

Structures of HIV protease ligands analyzed in this work. On the left is the structure of MK1 (indinavir), in the center 6EH, while the structure on the right is that of 6EF.

Figure 3

Calculated total energies for the HIV protease and MK1 (indinavir) system. (a) The data reported refer to a series of simulations (three replicates) for a water box containing 20,521 water molecules and the HIV protease. The line that best fits the data is also shown (the equation has the form $U_{tot}=401.6620T-284184.0648$, $R^2=0.9999$). (b) The data reported refer to a series of simulations (three replicates) for a water box containing 6172 water molecules and the MK1 molecule. The line that best fits the data is also shown (the equation has the form $U_{tot}=114.9582T-83678.3397$, $R^2=0.9999$). (c) The data reported refer to a series of simulations (three replicates) for a water box containing 20503 water molecules and the HIV protease/MK1 complex. The line that best fits the data is also shown (the equation has the form $U_{tot}=401.4182T-283721.8660$, $R^2=0.9999$).

Figure 4

Calculated total energies for the HIV protease and 6EH system. (a) The data reported refer to a series of simulations (three replicates) for a water box containing 19,870 water molecules and the HIV protease. The line that best fits the data is also shown (the equation has the form $U_{tot}=389.4320T-275541.9406$, $R^2=0.9999$). (b) The data reported refer to a series of simulations (three replicates) for a water box containing 5236 water molecules and the 6EH molecule. The line that best fits the data is also shown (the equation has the form $U_{tot}=97.4343T-71073.8775$, $R^2=0.9999$). (c) The data reported refer to a series of simulations (three replicates) for a water box containing 19853 water molecules and the HIV protease/6EH complex. The line that best fits the data is also shown (the equation has the form $U_{tot}=389.7589T-275415.0402$, $R^2=0.9999$).

Figure 5

Calculated total energies for the HIV protease and 6EF system. (a) The data reported refer to a series of simulations (three replicates) for a water box containing 19,883 water molecules and the HIV protease. The line that best fits the data is also shown (the equation has the form $U_{tot}=389.9753T-275808.8660$, $R^2=0.9999$). (b) The data reported refer to a series of simulations (three replicates) for a water box containing 5226 water molecules and the 6EF molecule. The line that best fits the data is also shown (the equation has the form $U_{tot}=97.4781T-70982.6109$, $R^2=0.9999$). (c) The data reported refer to a series of simulations (three replicates) for a water box containing 19,853 water molecules and the HIV protease/6EF complex. The line that best fits the data is also shown (the equation has the form $U_{tot}=389.5510T-275351.4252$, $R^2=0.9999$).