Page 104 - TD-3-4

Tumor Discovery Bioinformatics insights into CCL2 mutations

complexes were prepared for MD simulations using the of CCL2 proteins – indicators to classify protein residues –
GROMOS 43a1 force field. RMSD measures the average as well as changes in secondary structures between native
65
distance between the atoms of two superimposed molecular and mutant CCL2 proteins. 66
structures. It is commonly used in bioinformatics to assess
the similarity of 3D structures, particularly in the study 2.10. Docking of CCL2-native and -mutant proteins
of protein conformations. RMSD is calculated using the with receptor CCR2
61
formula: The objective of molecular docking for the ligand–receptor
RMSD = 1/ N ∑ i n =1 () iδ 2 (I) (CCL2–CCR2) binding process is to construct an accurate
binding free energy landscape and estimate the binding
affinity and binding kinetics. We introduced a novel hybrid
where δi represents the distance between the docking (HDOCK) protocol that integrates template-
corresponding atoms in the two structures, and N is the based and template-free approaches to enhance protein–
number of atoms considered. Typically, RMSD is calculated protein docking accuracy. Unlike traditional methods that
for the backbone heavy atoms or alpha carbon (Cα) atoms. focus on docking monomer proteins, HDOCK utilizes
RMSF is a measure of the average deviation of atomic homologous complex structures of the proteins involved.
positions from their mean positions over time in MD Initially, homologous complexes were identified using
simulations. It quantifies the fluctuations of individual BLAST searches of the PDB, with a sequence identity cutoff
atoms or residues, providing insights into the flexibility set at 25%. Among multiple hits, the optimal template was
and mobility of different regions within a protein structure. selected based on the highest sequence coverage, sequence
RMSF is calculated as the square root of the variance of the similarity, and resolution. This selected homologous
fluctuations around the average position using the formula: complex structure was then used as a docking template,
62
incorporating potential binding interface information. By
RMSF = (γi − ) i 2 (II) leveraging the structural information from the homologous
γ
complexes, HDOCK substantially improves the prediction
accuracy of protein–protein interactions. The individual
67
where γi represents the coordinates of atom i, and 〈γi〉 is
its ensemble average position. High RMSF values indicate protein structures are modeled using the homologous
complexes as templates, either by superimposing them onto
regions with considerable mobility, whereas low values available homologous complexes or through homology
suggest rigidity.
modeling. In our study, the individual component
68
To gain insights into the hydrogen bonding characteristics structures of the modeled complex were separated and
of the proteins, we analyzed the hydrogen bond geometries. subjected to a global sampling of potential binding modes
As the positions of hydrogen atoms are uncertain in these using our hierarchical protein–protein docking algorithm.
structures, they were excluded from the analysis. In total, This approach allows for a comprehensive exploration of
4195 hydrogen bonds were identified and categorized into possible interaction interfaces between the proteins. 69,70
four types based on the donor and acceptor atom origins: Symmetric docking was achieved by constraining the
main-chain to main-chain (MC-MC), main-chain to side- translational degrees of freedom to ensure symmetry
chain, side-chain to main-chain, and side-chain to side- along a specific axis. In cases where no complex structure
chain. The MC-MC hydrogen bonds were further classified templates were available, the individual protein structures
according to the secondary structures, including alpha for docking were modeled using their homologous
helices, beta strands, and non-secondary structures. monomer structures with Modeller. The binding modes
71
Furthermore, we investigated the correlations between obtained from the docking simulations were evaluated
hydrogen bond geometry and environmental factors, such and ranked according to their binding energy scores. To
as atomic depth, secondary structure, and partial charge. group similar conformations, an RMSD cutoff of 5.0 Å was
Atomic depth was defined as the distance to the nearest bulk applied for clustering. From the resulting clusters, the top
water molecule. This analysis provides a comprehensive 100 binding modes were examined and the best 10 models
understanding of the characteristics and determinants of were selected for submission to the Critical Assessment of
hydrogen bonding in high-resolution protein structures. Predicted Interactions initiative.
2.9. Examination of alterations in CCL2 protein 3. Results
secondary structure The results of this study demonstrated the considerable
The Database of Secondary Structure in Proteins tool impact of mutations in the CCL2 gene on the molecular
analyzes hydrogen bonding and other secondary structures stability, functionality, and structure of the resulting

Volume 3 Issue 4 (2024) 6 doi: 10.36922/td.3891