Simulations of AFPS in a continuum

As shown above the size of the explicit water simulations can be rather large, even for a medium sized protein as in the case of the sea raven antifreeze protein (113 amino acid residues and 5391 water). Simulations of that size can require a large amount of computer memory and disk space. If one is interested in the stability of a particular antifreeze protein or in general any protein and not concerned with the protein-solvent interactions, then an alternative method is available. In this case the simulation of a protein in which the explicit waters are represent by a structureless continuum. In this continuum picture the solvent is represented by a dielectric constant. This replacement of the explicit solvent model by a continuum is due to Bom and was initially used to calculate the solvation free energy of ions. For complex systems like proteins one uses the Poisson-Boltzmann equation to solve the continuum electrostatic problem. In 

Explicit Solvent, 300 K 3658 T1P3 Water PBC 90 X 36 X 36 9-11 A switched cutoff 60 ps heat, equil 80 ps data collection CHARMM code 

PB update every 10 steps Grid size 0.65 A 100 ps data collection UHBD code 

One final observation about the simulations on this protein should be made. As indicated in Table 2 the 300 K implicit simulation resulted in an unstable helix while the 273 K simulation yielded a stable helix. In fact the melting temperature of this helix is approximately 34 C. This brings up two main points. The first point is that the force-field parameters (CHARMM22) [28] used in the implicit solvent simulations to model protein structure are reasonable, however, the melting temperatures are a little bit too low. The second point is that in the implicit solvent simulations, small molecules as well as larger macromolecules can explore regions of conformational space that are not normally done in explicit solvent simulations. This fact is further supported in the Type III implicit solvent simulations. 

---