Premium
An efficient, differentiable hydration potential for peptides and proteins
Author(s) -
Augspurger Joseph D.,
Scheraga Harold A.
Publication year - 1996
Publication title -
journal of computational chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.907
H-Index - 188
eISSN - 1096-987X
pISSN - 0192-8651
DOI - 10.1002/(sici)1096-987x(199610)17:13<1549::aid-jcc6>3.0.co;2-s
Subject(s) - gaussian , chemistry , solvation , potential energy , atom (system on chip) , statistical physics , computational chemistry , van der waals force , thermodynamics , physics , mathematics , molecular physics , solvent , quantum mechanics , computer science , molecule , organic chemistry , embedded system
An approximate method for calculating the exposed volume of the hydration shell (VHS) about an atom, the R educed R adius I ndependent G aussian S phere (RRIGS) approximation, is presented. A key ingredient in this method is the use of reduced van der Waals radii so that the error of including only double overlap terms (and omitting multiple overlap terms) in calculating the VHS is balanced by a reduction in the magnitude of the double overlap terms. Also, the double overlap is modeled with a gaussian function. The RRIGS approximate calculation of the VHS is shown to be very accurate (the rms deviation of the VHS of each atom in avian pancreatic polypeptide and bovine pancreatic trypsin inhibitor was 14.0 and 15.8 Å 3 , respectively, out of a range of values between 0 and 600 Å 3 ). The RRIGS approximation is used to develop a potential function to represent the free energy of solvation for proteins. The pairwise gaussian form of the potential enables it to be incorporated into a gaussian representation of ECEPP (Empirical Conformational Energy Program for Peptides) for use in the Diffusion Equation Method (DEM) of global optimization. Inclusion of the effects of hydration by means of this potential is shown to require less than twice the computational time needed for computing the ECEPP conformational potential energy alone; this makes inclusion of solvent computationally feasible. Furthermore, this gaussian hydration potential function and its derivatives are continuous, so that it may be readily minimized. The combined potential of ECEPP/3 plus hydration is shown to require fewer energy evaluations per local minimization than ECEPP/3 alone for two small peptides. © 1996 by John Wiley & Sons, Inc.