Introduction
Protein-protein interactions (PPI) and protein-peptide interactions
(hereafter referred to as -PPepI) play a crucial role in many
cell-signaling and metabolic processes. In recent years, they have
become attractive targets for drug and biologics discovery [1-8]. To
interfere or modulate PPI, protein-protein complex may be disrupted or
interactions may be augmented by peptides or mimetics. Peptides have the
advantage of larger shelf life, less prone to proteolysis, feasibility
of their oral delivery and flexibility in optimization, screening and
synthesis, similar to small molecule based drug. Thus, to augment or
disrupt PPI by peptide-based therapeutic, it is worthwhile to understand
the differences between PPI and PPepI.
To characterize differences between PPI and PPepI, in the present work
we have focused on two important class of residues, namely, hotspot and
anchor residues. At the interface of molecular complex, certain residues
play a crucial role in governing such interactions. These residues
contributing most to the binding of molecular complexes are termed as
hotspot residues. As a rule of thumb, a hotspot residue is defined, if
its energy contribution to the binding of complex is more than 2
kcal/mol [9-10]. The binding affinity and specificity of the PPI and
PPepI are in essence governed by such energetically important hotspot
residues. Mutations at hotspots have been shown to cause dissociation of
complexes or disruption in the signaling processes [11].
Additionally, anchoring residue also play specific role in the initial
stage of molecular recognition. They avoid kinetically costly structural
rearrangements in the binding pathway allowing for a relatively smooth
recognition process [12]. Anchor residues are hotspot residues that
bury the highest amounts of solvent accessible surface area ( ≥ 100
Å2) upon binding [12-13].
To quantitatively probe the energetic contributions of residue to the
overall binding of molecular complexes, earlier experimental alanine
scanning mutagenesis were attempted [9]. This involves
experimentally mutating each residue at the interface to alanine and
then measuring the effect of the mutation on binding to the partner
protein or peptide. Extending the previous definition of hotspot more
specifically with the Ala-scanning method, a residue is considered a
hotspot residue, if its mutation to alanine gives rise to loss in
binding affinity ≥ 2 kcal/mol. Early on, limited data were generated
using experiment mutagenesis for many diverse complexes [14-22], and
few databases on experimentally determined hotspot residues in
protein-protein complexes such as AB-bind [14], ASEdb [15],
ATLAS [16], BID [17], DACUM [18], dbAMEPNI [19], dbMPIKT
[20], PROXiMATE [21] and SKEMPI 2.0 [22] have also been
developed. The identification of hotspot residues by experimental
techniques is a costly and time consuming job, which has led to the
development of a series of computational algorithms to predict hotspots
at the protein-protein interfaces [1, 23-39]. Serrano’s group have
developed FoldX based on an empirical potential for rapid evaluation of
effect of mutations in proteins and nucleic acids [23,24]. Flex_ddG
method in Rosetta uses a combination of sophisticated Monte Carlo
sampling, minimization, and specialized force fields [25-27].
BeAtMuSiC is a tool, which employ statistical potentials adapted to a
coarse-grained representation of protein structures [28] and mCSM
(mutation Cutoff Scanning Matrix) is based on graph-based structural
signatures, which encodes distance patterns between atoms to represent
protein residue environments [29]. Various other methods and their
applications have also been discussed in the literature [30-36,
44-45].
In the present work, we have used Bioluminate Residue scanning approach
implemented in Schrödinger [37-38]. The method uses implicit
solvation based MM-GBSA (Molecular Mechanics with Generalized Born and
Surface Area) with the OPLS2005 force field, VSGB2.0 solvent model and
rotamer search algorithms from Prime [37-38]. To understand the
differences in the nature of hotspots in protein-protein and
protein-peptide complexes using benchmarking data sets, a comprehensive
Ala scanning, comprising of over 5500 mutation analyses were performed.
Materials and Methods