pathways between allosteric sites and other active sites including
catalytic sites or ligand-binding sites also needs to be established.
Computational approaches including both sequence and
structure-based methods have been widely used to investigate the
molecular basis of allosteric regulation and communication [7–
11]. Sequence-based models can describe allosteric sites in terms
of conservation and coevolution properties and enumerate potential communication pathways by constructing evolutionary networks [12]. Network-based structural studies (also called PSN)
have also demonstrated that the topology and connectivity of protein structures provide a robust framework for understanding allosteric effects in terms of local and global graph-based parameters
[13]. Recent evidence supports the view that allosteric communication is facilitated by the intrinsic dynamics of the biomolecules
[14]. These methods, combined with molecular dynamics (MD)
simulations, have been recently applied to elucidate allosteric communication pathways of diverse protein systems [15–19]. In the
most recent studies, Verkhivker et al. [20, 21] have combined
evolutionary analysis, MD simulations of the molecular chaperone
of Hsp90 with the network analysis, and perturbation response
scanning (PRS) approach to probe key sites in allosteric
mechanisms.
An alternative dynamical approach, Elastic network model
(ENM) [22], was first introduced by Tirion to study the intrinsic
dynamics of proteins by using normal mode analysis. In ENMs, a
protein structure was considered as a network consisting of a set of
residues interconnected by elastic springs. Currently, two main
types of ENMs are widely used, which are the Gaussian network
model (GNM) [23] and the anisotropic network model (ANM)
[24]. Based on ENMs, several theoretical approaches have been
made toward understanding the molecular basis of allosteric communication. ENM combined with dynamics perturbation analysis
(DPA) [25] and PRS [26, 27] were developed to detect the key
residues whose perturbation couple structural dynamic changes at
distal distance. By using ENM to calculate the correlations between
the fluctuations in spring length of pairwise residues, a new method
was also proposed to identify allosteric residues [28]. In addition, a
thermodynamic method based on ENM was proposed to predict
the allosteric sites on the protein surface [29]. ENM and PSN are
two coarse-grain methods, and the integration of them may help to
understand the pathways of communication between spatially distant sites [30]. The first attempt to combine ENM and PCN has
been reported to investigate allosteric communication pathways in
the PDZ2 domain [31]. The successful applications of combining
ENM and PSN in the prediction of structural dynamics and allosteric sites have been proved in several protein systems [32] and
bacterial ribosome [33].
22
Guang Hu
catalytic sites or ligand-binding sites also needs to be established.
Computational approaches including both sequence and
structure-based methods have been widely used to investigate the
molecular basis of allosteric regulation and communication [7–
11]. Sequence-based models can describe allosteric sites in terms
of conservation and coevolution properties and enumerate potential communication pathways by constructing evolutionary networks [12]. Network-based structural studies (also called PSN)
have also demonstrated that the topology and connectivity of protein structures provide a robust framework for understanding allosteric effects in terms of local and global graph-based parameters
[13]. Recent evidence supports the view that allosteric communication is facilitated by the intrinsic dynamics of the biomolecules
[14]. These methods, combined with molecular dynamics (MD)
simulations, have been recently applied to elucidate allosteric communication pathways of diverse protein systems [15–19]. In the
most recent studies, Verkhivker et al. [20, 21] have combined
evolutionary analysis, MD simulations of the molecular chaperone
of Hsp90 with the network analysis, and perturbation response
scanning (PRS) approach to probe key sites in allosteric
mechanisms.
An alternative dynamical approach, Elastic network model
(ENM) [22], was first introduced by Tirion to study the intrinsic
dynamics of proteins by using normal mode analysis. In ENMs, a
protein structure was considered as a network consisting of a set of
residues interconnected by elastic springs. Currently, two main
types of ENMs are widely used, which are the Gaussian network
model (GNM) [23] and the anisotropic network model (ANM)
[24]. Based on ENMs, several theoretical approaches have been
made toward understanding the molecular basis of allosteric communication. ENM combined with dynamics perturbation analysis
(DPA) [25] and PRS [26, 27] were developed to detect the key
residues whose perturbation couple structural dynamic changes at
distal distance. By using ENM to calculate the correlations between
the fluctuations in spring length of pairwise residues, a new method
was also proposed to identify allosteric residues [28]. In addition, a
thermodynamic method based on ENM was proposed to predict
the allosteric sites on the protein surface [29]. ENM and PSN are
two coarse-grain methods, and the integration of them may help to
understand the pathways of communication between spatially distant sites [30]. The first attempt to combine ENM and PCN has
been reported to investigate allosteric communication pathways in
the PDZ2 domain [31]. The successful applications of combining
ENM and PSN in the prediction of structural dynamics and allosteric sites have been proved in several protein systems [32] and
bacterial ribosome [33].
22
Guang Hu