The Journal of Organic Chemistry
Article
Figure 5. Distribution of Z/E isomers of NMe-thioamide 3b obtained by REST simulations at 300 K in various solvents. Horizontal axis is the
dihedral angle of the thioamide rotation: conformers between −50° and +50° are regarded as the Z conformer and conformers between −150° to
−180° and 150° to 180° are regarded as the E conformer. The vertical axis is the frequency of the conformer (%).
conjugation will be enhanced (see Bond Model Analysis
below).
MD based simulations clearly reproduce the experimental
solvent effects on the Z/E ratio of the NH-thioacetamide.
On the other hand, similar REST simulations reproduced
the lack of a significant solvent effect on the Z/E ratio in the
case of the N-Me-thioacetamide, as well as the Z biases (Figure
5). The calculations reproduced the experimental bias of Z/E
distribution of the N-Me thioamide 3b: in CHCl3: simulated
value (Z/E): 99.3:0.7; experiment: >99:<1; in DMSO:
simulated value(Z/E): 98.3:1.7; experiment: > 99:<1.
Radial Distribution Function (g(r)). The interactions
between the thioamides and the solvent can be analyzed
from the trajectories of MD simulations obtained at 300 K for
40 ns (force field: OPSL3e) by means of the radial distribution
function (g(r)) (Figures 6 and 7),20 which expresses the
probability distribution of the distance between the two
molecules in solution. The target atom in one molecule is
placed at the origin, and the density of the target atom in the
other molecule at the distance r is divided by the density of the
bulk system. The horizontal axis is the distance between the
atoms (Å), the left vertical axis is the radial distribution
function g(r) (blue line) with the ratio of the bulk solvent with
solvent structure set to 1.0, and the right vertical axis (Figures
6 and 7) is the integrated number of solvents (red curve,
coordination number), indicating how many molecules are
found in the range of each coordination sphere.
1. Polar and Hydrogen-Bond-Donating Solvents Such as
DMSO, Acetone, and MeOH. In polar and hydrogen-bond-
donating solvents such DMSO, acetone, and protic MeOH, the
oxygen atom of the solvent molecule can coordinate to the
acidic NH proton of the NH-thioamide 3a,21 giving a
distribution starting at 1.8 Å with a peak at approximately
2.1 Å for both Z and E thioamides ((3a)NH---O(DMSO or
acetone or CH3OH) in Figure 6a−c). From the volume
integration of g(r) in the range of 1.8−3.5 Å, approximately
one molecule of the solvent interacts with the thioamide NH
atom. The values of the radial distribution function (g(r)) and
the coordination number of solvents at the first peak are
smaller in the case of Z thioamide than those in the E
thioamide (Figure 6). This is probably due to steric hindrance
around the NH group of the Z thioamide conformer, such as
the sulfur atom of the thioamide, which hampers access of
solvent molecules to the NH moiety. This difference in solvent
coordination between Z and E thioamide conformers may be
attributed at least in part to the solvent dependency of the
stability of these conformers: these three solvents, DMSO,
4. Computational Study Based on Molecular Dynam-
ics. In order to understand the roles of solvents, acetyl alkyl
substituents, and N-methylation, we carried out molecular
dynamics (MD) calculations, including explicit solvent
molecules. Based on the behaviors of solvents found in the
MD simulations, we carried out DFT calculations that included
one or two explict solvent molecules in an implicit solvent
model, which can mimic the more realistic solvation
environment found in the MD calculations. These calculations
revealed hydrogen bonding mainly between the thioamide NH
proton and the solvent heteroatom(s) in polar solvents, and
also identified a weak interaction of halogenated solvents
(CHCl3 and CH2Cl2) with the thioamide (NH−CS)
functionality.
Replica Exchange with Solute Tempering (REST) Simu-
lations. To investigate the solvent dependence of thioamide
isomerization, we performed accelerated molecular dynamics
simulations of 3a (p-Me) using the replica exchange with
solute tempering (REST) method18 with the OPSL3e force
field in various solvent systems (Figure 4; details of simulation:
see the Experimental Section). The OPSL3e force field is
improved by additional parametrization and becomes appli-
cable to many organic compounds.19 The Z/E ratio of NH-
thioamide 3a was simulated in various solvents at 966 K. The
populations of Z and E conformations were calculated by using
1,000 snapshot structures in the trajectories at 300 K, obtained
in the REST calculations in a different solvent, over the range
of 300 to 966 K. In CHCl3, the Z/E ratio was simulated to be
40:60 [experiment: 49:51] at 300 K. In the case of DMSO, the
Z/E ratio was simulated to be 4:96 [experiment: 8:92]. Other
solvent systems gave the following simulated values: CH2Cl2:
35:65 [experiment: 36:64] (Z/E); acetone: 12:88 [experiment:
11:89] (Z/E); methanol: 12:88 [experiment: 9:91] (Z/E).
Therefore, the REST calculations are in good agreement with
the experimental trend in the Z/E ratio of the thioamide 3a.
The reduction of the Z isomer ratio in polar solvent from
that in nonpolar solvent was also reproduced in the REST
calculations for another NH-thioamides 1a (p-H): in CHCl3:
simulated value (Z/E): 37:63; experiment: 45:55; in CH2Cl2:
simulated value (Z/E): 25:75; experiment: 33:67; in acetone:
simulated value(Z/E): 21:79; experiment: 9:91; in methanol:
simulated value (Z/E): 14:86; experiment: 9:91; in DMSO:
simulated value(Z/E): 8:92; experiment: 8:92. Therefore, the
G
J. Org. Chem. XXXX, XXX, XXX−XXX