Γξx¼Cξ
CξxDECξ(^) ð 5 Þ
where Cξ
PN
i¼ 1 CξðÞi=N
is the average network centrality,
and Cξx
DE
is a value evaluated for a newly constructed network
when the nodexis removed from the original network.
The idea of network vulnerability is similar to point mutagene-
sis assay, which measures the effect of mutations up to the degree
where proteins can retain their activity. Adopting this vulnerability
concept, we performed in silico glycine scanning experiments of the
constructed residue interaction network of A2AAR. As straightfor-
wardly implicated by the term “glycine scanning,” the side chain of
each residue was deleted systematically, and the corresponding
residue interaction network was reconstructed. Then, we compared
the original residue interaction network to the reconstructed one to
measure how much the centralities of whole network was changed
when the side chain of specific amino acid was deleted. The main
difference between our in silico glycine scanning analysis and the
previous studies applying network analysis [38, 40] is that only side
chain atoms were deleted, rather than the extraction of entire node
in the network. Therefore, our method could be considered more
realistic as it mimics changes similar to the experimental mutation
studies.
We assessed the effect of deleting side chains by calculating the
changes in average betweenness centralities (hCBi). In our case,
readjustment of local environment due to the removal of side
chain is not considered. The primary aim of this analysis is to
quantitatively assess the role of a specific residue side chain in the
original residue interaction network. The greater the role played by
the deleted side chain in maintaining the network structure, the
more significant the response of average network centrality would
be to the removal of that particular residue. Interestingly, the
residues with high network vulnerabilities included the regions
around NPxxY motif in TM7 and proline kink in TM6, which
were reported as important residues providing structural con-
straints in GPCRs [41, 42]. Along with our analysis, these regions
seem to be potent for maintaining the integrity of whole residue
interaction network modulating the entire activation process.
2.2.5 Cross-Correlation
Between Regions Based
on the Structural
Ensembles
In the active form of GPCR structure, the ligand binding site in the
extracellular side and the intracellular G-protein binding region are
functionally coupled, and this coupling is mediated by a structural
reorganization of 7TM helices. To quantify such long-range cou-
pling in the dynamics of A2AAR, we calculated cross-correlation
between residues (CCij) in terms ofCB(Eq.4) by using the
464 Shaherin Basith et al.