Section 4: Compute binding enthalpies within the APR scheme
The Amber tutorial for computing binding enthalpy using a multi-box approach can be found here: TUTORIAL A21: Computing Binding Enthalpy Values, by Andrew T. Fenley & Michael K. Gilson. Binding enthalpy calculations have long been considered problematic, due to the large fluctuations in potential energies of MD simulations. However, it has been shown that for host-guest systems, numerically precise binding enthalpies can be computed with microsecond time-scale simulations readily completed on GPUs. (Fenley et al.)
Binding enthalpies can also be computed conveniently within the APR framework, but they are not automatically computed at the moment. Compared with the multi-box approach, in which four sets of simulations are needed – the free host, the free guest, the complex and pure water – the technique of computing binding enthalpies within the APR scheme, also called a single-box approach, computes binding enthalpies by simply extending the simulation length of the first window (a00, i.e. the bound state) and last window (p45, i.e. the unbound state) to the microsecond scale, and then computing the difference of the potential energies between those two states. You can use the restart file generated from the equilibration phase in these windows and increase the number of simulation steps in the mdin file. Note that the potential energies should be extracted from the "EAMBER" entry in the Amber output file for each frame, for which the non-restraint energy is recorded, instead of "EPtot". The results of our enthalpy simulations for the OA-cba system are listed below:
The experimental binding enthalpy of OA – 4-cyanobenzoic acid complex is - 4.39 ± 0.03 kcal/mol (See Sullivan et al.). The results you compute should agree within our enthalpy estimate within the range of uncertainties, if equal length or longer simulations are conducted. The results computed by multi-box and single-box approach for the same host-guest complex should also match each other within uncertainties.