A method for obtaining electrostatic multipoles for protein-ligand systems is proposed. Atomic multipoles (up to octupoles) are calculated with GAMESS using the Distributed Multipole Analysis (DMA) method. The multipoles are subsequently used for the Coulomb term in our new intermolecular force field (IEFF).
IEFF is used for the estimation of protein-ligand interaction energies and ligand poses. The protein is fragmented into either capped mono-residue or tripeptides. Using 12-20 residues model systems, we compare both mono-residue/tripeptide approximations to the entire protein calculation.
We show that the tripeptide approximation for calculating atomic multipoles is significantly more reliable for the estimation of protein-ligand interaction energy, while the differences between mono-residue/tripeptide are smaller for pose predictions.
The usage of the proposed potential is illustrated for both hydrophobic and water mediated binding.
Stan Wlodek, Senior Scientist
OpenEye Scientific Software
Stan Wlodek and Laszlo Fusti-Molnar
A method for obtaining electrostatic multipoles for protein-ligand systems is proposed. Atomic multipoles (up to octupoles) are calculated with GAMESS using the Distributed Multipole Analysis (DMA) method. The multipoles are subsequently used for the Coulomb term in our new intermolecular force field (IEFF).
IEFF is used for the estimation of protein-ligand interaction energies and ligand poses. The protein is fragmented into either capped mono-residue or tripeptides. Using 12-20 residues model systems, we compare both mono-residue/tripeptide approximations to the entire protein calculation.
We show that the tripeptide approximation for calculating atomic multipoles is significantly more reliable for the estimation of protein-ligand interaction energy, while the differences between mono-residue/tripeptide are smaller for pose predictions.
The usage of the proposed potential is illustrated for both hydrophobic and water mediated binding.
Stan Wlodek received his Ph.D. in Physical Chemistry from the University of Warsaw. He spent several years in Canada as a post-doc and then went south to the University of Houston where he worked with Andy MacCammon. He moved to Santa Fe in 1998. In the fall of 2001, Stan started working as a consultant with OpenEye, implementing the AM1-BCC charging method of Christopher Bayly. He decided against taking the coaching job for Poland in the 2002 World Cup, which can be directly related to Poland's lethargic performance, and joined OpenEye. Notwithstanding on how well Poland does in the 2010 competition, we anticipate Stan remaining at OpenEye and continuing his work on SZYBKI (he made us name it that) and AFITT (our bad).