Can the Structure of the Hydrophobic Core Determine the Complexation Site? - Identification of Ligand Binding Site and Protein-Protein Interaction Area - page 49

presented model: “antifreeze” proteins (DeVries and Wohlschlag 1969 ; Jarov et al.
2004 ) and “downhill” (fast-folding) proteins (Fisher and DeLisa 2008 ; Dyer 2007 ;
Ozkan et al. 2002 ; Zhu et al. 2003 ). Both groups confirm the predictions of the
model (Banach et al. 2012 ; Roterman et al. 2011 ) .
Fast-folding (or “downhill”) proteins have been experimentally proven to possess
the ability to undergo rapid and reversible folding in vitro . This property suggests
high spontaneity of the folding process, with little reliance on external conditions.
Conformance with theoretical predictions was assessed on the basis of distance
entropy values (O/T and O/R) (Banach et al. 2011 ) .
Analyzing examples of structurally accordant proteins validates our model by
confirming that such proteins do indeed exist. The physiochemical properties of
these proteins (“fast-folding” group) suggest that their structure is only affected by
the aqueous environment. Thus, a model acknowledging the relationship between the
polypeptide chain and the aqueous environment seems sufficient to determine the
structural ordering of protein molecules.
When discussing antifreeze proteins, the influence of mutations should be taken
into account (Banach et al. 2012 ). The PDB database usually lists several mutations
per protein. Analysis of the hydrophobic core of 1MSI (Jia et al. 1996 ) indicates that
while most mutations do not significantly alter the structure of this protein, specific
mutations at position 16 (A16M, A16T, A16M, A16C, A16R, A16Y) result in reor-
ganization of the hydrophobic core in a way which breaks conformance with our
model. This can have far-reaching implications for the shape of the entire protein
molecule and for its biological properties (Banach et al. 2011 ) .
Having presented a selection of proteins whose structure follows the presented
model we should devote our attention to the observed discrepancies. Among
enzymes with well-defined active sites hydrolases appear to exhibit particularly
good agreement with the “fuzzy oil drop” model. Studying their

Δ H pro fi les points
to specific areas where selected amino acids diverge from the model in terms of
hydrophobicity. These amino acids correspond to sites of enzymatic activity, i.e. the
binding pockets (which translate into Δ H pro fi le maxima).

From among many classes of enzymes, our model is particularly efficient in
predicting the active sites of hydrolases (Brylinski et al. 2007b, c ) - this is why
hydrolases have been singled out for in-depth analysis, which is presented in
(Prymula et al. 2011 ) .

3.3 Summary

In summarizing the presented work we should state that irregularities in the structure
of the protein's hydrophobic core, triggered e.g. by the presence of a ligand, provide
a good starting point for identification of active sites (ligand-binding and protein
complexation areas). These irregularities correspond to minima or maxima of the

Δ H hydrophobicity profile, which, in turn, indicates which residues are in direct
contact with a ligand or with another protein molecule. Measuring deformations of