W9: Computational modeling

• testing hypothesized reaction mechanisms

• discussing the origin of catalysis in an enzyme

  • e.g. LBHBs

• identifying intermediates

• help interpret crystal data

Theoretically, the fundamental laws to describe chemical reactions are known and we should be able to describe them.

Practically, real systems are so complex (number of interactions) that we do not have the computational power to describe them in a reasonable time.

—> approximations

Molecular mechanics:

• model: atoms behave like balls connected with springs

• bonds cannot break!

• good way to describe conformational changes

(Molecular mechanics uses harmonic oscillator approximation (black), therefore, no bond breakage possible)

Quantum mechanics:

• model: atoms have nuclei and electrons

• more precise but still an approximation

• approach to describe a molecule: combine wave functions of orbitals, minimize energy (that is your bond)

• problem: even equations with relatively few atoms can take a long time to be solved

  • only used for up to a few hundred atoms

To describe proteins/enzyme mechanisms, both approaches are combined (MM for conformational changes, movements of substrates, ions, water, QM for the actual catalysis)

bond stretching: 1 bond

angle bending: 2 bonds involved

torsion: 3 bonds involved

non-bonded terms: describe interaction of atoms that are several bonds apart

• electrostatic interactions

• vdW interactions

During computational simulations of chemical reactions, periodic boundary conditions are used to restrict the area you have to consider.

The theoretical reaction vessel is divided into cubes that contain the relevant molecules.

Molecules can leave the space but enter at the same time from the opposite direction.

• difficult to describe coordination

• d-orbitals of transition metals are difficult to calculate

• exchange of ligands (bond breakage) cannot be modelled