Systematic Correlation of Structural, Thermodynamic and Residual Solvation Properties of Hydrophobic Substituents in Hydrophobic Pockets Using Thermolysin as a Case Study
Water molecules participate besides protein and ligand as an additional binding partner in every in vivo protein–ligand binding process. The displacement of water molecules from apolar surfaces of solutes is considered the driving force of the hydrophobic effect. It is generally assumed that the...
|PDF Full Text
No Tags, Be the first to tag this record!
|Water molecules participate besides protein and ligand as an additional binding partner in
every in vivo protein–ligand binding process. The displacement of water molecules from
apolar surfaces of solutes is considered the driving force of the hydrophobic effect. It is
generally assumed that the mobility of the water molecules increases through the
displacement, and, as a consequence, entropy increases. This explanation, which is based on
experiments with simple model systems, is, however, insufficient to describe the hydrophobic
effect as part of the highly complex protein–ligand complex formation process. For instance,
the displacement of water molecules from apolar surfaces that already exhibit an increased
mobility before their displacement can result in an enthalpic advantage. Furthermore, it has
to be considered that by the formation of the protein–ligand complex a new solvent-exposed
surface is created, around which water molecules have to rearrange. The present thesis
focuses on the impact of the latter effect on the thermodynamic and kinetic binding
properties of a given ligand.
A congeneric ligand series comprised of nine ligands binding to the model protein
thermolysin (TLN) was analyzed to determine the impact of the rearrangement of water
molecules around the surface of a newly formed protein–ligand complex on the
thermodynamic binding properties of a ligand. The protein–ligand complexes were
characterized structurally by X-ray crystallography and thermodynamically by isothermal
titration calorimetry (ITC). The only structural difference between the ligands was their
strictly apolar P2’ substituent, which changed in size from a methyl to a phenylethyl group.
The P2’ group interacts with the flat, apolar, and well-solvated S2’ pocket of TLN. Depending
on the bound ligand, the solvent-exposed surface of the protein–ligand complex changes. The
ITC measurements revealed strong thermodynamic differences between the different ligands.
The structural analysis showed ligand-coating water networks pronounced to varying
degrees. A pronounced water network clearly correlated with a favorable enthalpic and less
favorable entropic term, and overall resulted in an affinity gain.
Based on these results, new P2’ substituents were rationally designed with the aim to achieve
stronger stabilization of the adjacent water networks and thereby further increase ligand
affinity. First, the quality of the putative water networks was validated using molecular
dynamics (MD) simulations. Subsequently, the proposed ligands were synthesized, crystallized in complex with TLN, and analyzed thermodynamically. Additionally, a kinetic
characterization using surface plasmon resonance (SPR) was performed. The
crystallographically determined water networks adjacent to the P2’ substituents were in line
with their predictions conducted by MD simulations. The ligands showed increasingly
pronounced water networks as well as a slight enthalpy-drive affinity increase compared to
the ligands from the initial study. The ligand with the highest affinity showed an almost
perfect water network as well as a significantly reduced dissociation constant.
To analyze the influence of the ligand-coating water networks on the kinetic binding
properties of a ligand, seventeen congeneric TLN ligands exhibiting different P2’ groups were
kinetically (by SPR) and crystallographically characterized. The different degree of the water
network stabilization showed only a minor influence on the binding kinetic properties. By
contrast, the strength of the interaction between the ligand and Asn112 proved crucial for the
magnitude of the dissociation rate constant. A strong interaction resulted in a considerably
prolonged residence time of the ligand by hindering TLN to undergo a conformational
transition that is necessary for ligand release.
In the last study, the reason for the exceptionally high affinity gain for addressing the deep,
apolar S1’ pocket of TLN with apolar ligand portions was investigated. Therefore, a
congeneric TLN ligand series substituted with differently large apolar P1’ substituents
(ranging from a single hydrogen atom to an iso-butyl group) was analyzed. The exchange of
the hydrogen atom at the P1’ position with a single methyl group already results in a 100-fold
affinity increase of the ligand. To elucidate the molecular mechanism behind this
considerable affinity gain, the solvation state of the S1’ pocket was carefully analyzed. The
results strongly indicate that the S1’ pocket is completely free of the presence of any water
molecules. Thus, the huge affinity gain was attributed to the absence of an energetically costly
The data presented in this thesis show that to describe the thermodynamic signature of the
hydrophobic effect it is necessary to explicitly consider the change of the thermodynamic
properties of every involved water molecule. Solely considering the buried apolar surface area
and assigning an entropic term to it is not sufficient. The increasing stabilization of the water
network adjacent to the protein-bound ligand represents a promising approach — quite
independent of specific properties of the target protein — to optimize the thermodynamic
profile of a given ligand. This approach also allows fine-tuning of the kinetic binding