Polarizable Embedding Potentials through Molecular Fractionation with Conjugate Caps Including Hydrogen Bonds

Polarizable embedding (PE) is an advanced fragment-based classical embedding model similar to quantum mechanics/molecular mechanics (QM/MM). Unlike mechanical or electrostatic embedding, the polarization between the quantum and classical regions is reciprocal in PE. The quality of the embedding potential is critical to provide accurate results for both spectroscopic properties and dynamical processes. Typically, the embedding potential is derived from the fragmentation of the classical region into smaller fragments. From each individual fragment, the potential parameters, i.e., multipoles and polarizabilities, are derived based on ab-initio calculations. For solvents and other small molecules, the fragments consist of the individual molecules, while for larger molecules (e.g., proteins), further fragmentation is needed. A sufficiently robust fragmentation scheme is one of the most essential key elements for an accurate embedding potential. One such fragmentation strategy is the molecular fractionation with conjugate caps (MFCC) used in the PE model. As is widely known, hydrogen bonds play a key role in many biomolecular systems, e.g., in proteins, where they are responsible for the secondary structure. In this work, we assess the effects of including hydrogen-bond fragmentation in the MFCC procedure (MFCC-HB) for deriving the multipoles and polarizabilities. It is implemented in PyFraME, a Python package that automatizes the generation of embedding potentials. The MFCC-HB extension is evaluated on several molecular systems, ranging from small model systems to proteins, both directly in terms of electrostatic and embedding potentials and indirectly in terms of selected spectroscopic properties of chromophores embedded in water and in complex protein environments.