Proteins: Ssp DnaE Intein - Multiscale Modeling in Biomechanics and Mechanobiology - page 9

The other method that may be used to add solvent effects to DFT calculations is
through an implicit solvent method. In this method, the solvent effects are simulated
through dielectric fields rather than explicit solvent molecules. Using an implicit
solvent method is advantageous due to the fact it side steps the necessity for a large
number of solvent molecules as well as a description of their orientation within the
system and instead, the solvent environment is determined by the value of the dielec-
tric constant. However, although the computational cost is not as extreme as including
large numbers of solvent molecules, the calculations are still relatively expensive and
more time than standard DFT calculations is required to perform them. For our sys-
tem in question, we used the Polarized Continuum Model (PCM) [ 37 ] to simulate
the presence of water as would be in experiment. This method encloses the system
in a block of dielectric material, where withinspheres are carved out with a radius
dependent on the type of atom being enclosed. Care must also be taken with solvent
calculations in that the accuracy must be increased beyond default parameters to
ensure proper values are obtained. Our calculations were performed first in vacuum
( 1) and geometries obtained were used in single point calculations involving

solvents. We believe this is an acceptable method to help curb computational cost
since tests showed that the energy difference between the solvent single point cal-
culations using the vacuum optimized geometries and geometries obtained through
optimizations in solvent were negligible. Furthermore, two solvents were considered.
Initially, we performed calculations in full water solvent (

ʵ =

78.3553), however after
a literature search, we found that the dielectric constant within biological systems is
generally less than that of the external solvent experimentally [ 38 , 39 ]. Instead, when
water is the solvent used in experiment, the dielectric constant is generally between
ʵ =

ʵ =

15 in the interior part of the system. As such, we performed calculations
with dielectric constant

4 −

ʵ = 10. Our results for the reaction barrier study can be

found in Table 1.2 .
Examining the table, our values show that the non-allowed conformation provides
a lower barrier and thus reaction rate when threonine is present. However, in the case
of alanine, the allowed conformation would provide a lower reaction rate than the non-
allowed. Comparing between threonine and alanine, our results suggest that alanine
should greatly slow the reaction rate of the N-S acyl shift and although our results
represent an upper bound due to the lack of thermal dependence in our calculations,

Table 1.2 Values obtained from the reaction barrier calculations

ˆ ˈ Region Barrier (eV) Charge (e)

S Cys O Gly

Thr − 124.61 52.95 A 0.6 − 0.37 − 0.65

− 123.19 77.81 NA 1.5 − 0.26 − 0.67

Ala − 119.64 50.68 A 1.1 − 0.65 − 0.67

0.68
S Cys and O Gly are the sulfur on the catalytic cysteine and oxygen of the carbonyl group on glycine,
respectively

− 123.19 77.81 NA 2.1 − 0.77 −