Improving Generalized Born Models by Exploiting Connections to Polarizable Continuum Models. II. Corrections for Salt Effects

Abstract

A previous analytical investigation of the generalized Born (GB) implicit solvation model is extended to solvents of nonzero ionic strength. The GB model with salt effects (GB-SE) is shown to resemble the Debye-Hückel-like screening model (DESMO), a polarizable continuum model (PCM) that we have recently developed for salty solutions. DESMO may be regarded either as a generalization of the conductor-like PCM (C-PCM) that extends C-PCM to electrolyte solutions or alternatively as a generalization of Debye-Hückel theory to arbitrary cavity shapes. The connection between GB-SE and DESMO suggests how the former can be modified to account for the exclusion of mobile ions from the cavity interior, an effect that is typically absent in GB-SE models. We propose two simple GB-SE models that are exact for a point charge in a spherical cavity and that introduce the ability to account, albeit approximately, for the finite size of the mobile ions. The accuracy of these new models is demonstrated by applications to both model systems and real proteins. These tests also demonstrate the accuracy of the DESMO approach, as compared to more sophisticated PCMs developed for electrolyte solutions.

Including implicit solvation in the bond capacity polarization model

“…The coupling provides a way to include polarization in the solvent (and implicitly in the solute) without including a large number of explicit solvent molecules, and this may be a way to retain high accuracy in terms of solvation energies without a large reduction in computational efficiency. 40 The C-PCM model, on the other hand, is a continuum model, where all solute charges interact with all the surface charges. The exact definition of the cavity surface depends on the specific model 7,9 but will usually be allowed to change if the solute structure changes.…”

“…(8). 40 Equations 10and (13) involve the inversion of AGB and A † GB having the same (N, N) dimension as the original kernel ABC. The increase in computational cost by coupling the GB model to the BC equations is thus expected to be very small.…”

Dielectric continuum methods for quantum chemistry

“…113,149 It has been suggested that the "optimal" atomic radius for a given atom likely ought to vary as a function of its partial atomic charge, 150,151 and probably as a function of the solvent's dielectric constant as well. 94,132,152,153 In quantum chemistry, the latter effect is generally neglected whereas the former is handled solvent probe sphere solvent-accessible surface (SAS) van der Waals (vdW) surface solvent-excluded surface (SES) vdW F I G U R E 2 Various constructions of the solute cavity surface, using a set of atomic spheres (in gray) whose envelope defines the van der Waals (vdW) surface, shown in black. The solvent-accessible surface (SAS, in green) is defined either by augmenting the atomic radii by a probe radius (R A = R vdW,A + R probe ) or equivalently as the center point of the probe sphere as it rolls over the vdW surface.…”

“…This procedure effectively forces the pairwise GB ansatz in Equation (4.4) to reproduce the exact electrostatic energy defined by the continuum electrostatics problem, and the values of R A determined in this way have been called the "perfect" Born radii. 419,423 However, the procedure just described is not a practical one, although it can be used to generate data sets of exact pairwise interactions, 23,152,153 which might suggest new analytical forms to replace Equation (4.5). 23,152,153,425,426 For practical purposes the radii R K f g are usually determined using surface integration procedures, the most popular of which is 418,419,422,424 R…”

Twenty‐five years of nucleic acid simulations

“…A great deal of effort has gone into testing variants of these ideas against both numerical Poisson calculations and against experiment . Extensions have been made to include salt effects at the Debye‐Huckel level of approximation …”