consistent with the opening motion of β 1 and β 2 in opposite directions, while two α chains are relatively stable, mainly in a rigidbody-type motion. This observation is consistent with the pervious
molecular dynamics study, which has shown that β chains are more
strongly linked to the quaternary transition than α chains [45].
ANM Mode 2.
The second ANM mode of the T-Hbs (also the first ANM mode of
the R-Hb) shows another global motion involving quaternary
changes of two dimers (Fig. 5c), namely, α 1 β 1 dimer, exhibiting a
torsional rotation in an opposite direction with α 2 β 2 dimer, coordinated by the hinges at the α 1 -β 2 and β 1 -α 2 interfaces. This hingebending rotation defines the intrinsic dynamics of Hbs, which may
facilitate their allosteric communication pathways via hinge regions.
4 Notes
1. Main advantages of ENMs lie at its simplicity, whose modes at
low frequencies are enough to capture intrinsic dynamics and
allosteric properties of biomolecular systems. Thus, ENM
methods are applicable to large systems, as well as for highthroughput investigation of protein data.
2. There are some limitations in ENMs since ENMs only consider
Cα atoms without considering specific interactions. It can be
known that allostery should take into account the roles of
flexible regions such as loops. ENMs may not be suitable for
modeling the allosteric effects of these flexible regions.
3. Additionally, there are some other normal mode analysis-based
web servers to predict allosteric sites and signal propagation
pathways in proteins and their complexes, such as PARS [46],
SPACER [47], AlloPred [48], and DynOmics [49] servers.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (31872723) and a Project Funded by the Priority
Academic Program Development (PAPD) of Jiangsu Higher Education Institutions.
The author thanks Prof. Ivet Bahar for giving the opportunity
to study Elastic network models and ProDy in her lab. The author
also thanks Drs. Hongchun Li and Chakra Chennubhotla for
providing programming codes.
Identification of Allosteric Effects in Proteins by Elastic Network Models
33
molecular dynamics study, which has shown that β chains are more
strongly linked to the quaternary transition than α chains [45].
ANM Mode 2.
The second ANM mode of the T-Hbs (also the first ANM mode of
the R-Hb) shows another global motion involving quaternary
changes of two dimers (Fig. 5c), namely, α 1 β 1 dimer, exhibiting a
torsional rotation in an opposite direction with α 2 β 2 dimer, coordinated by the hinges at the α 1 -β 2 and β 1 -α 2 interfaces. This hingebending rotation defines the intrinsic dynamics of Hbs, which may
facilitate their allosteric communication pathways via hinge regions.
4 Notes
1. Main advantages of ENMs lie at its simplicity, whose modes at
low frequencies are enough to capture intrinsic dynamics and
allosteric properties of biomolecular systems. Thus, ENM
methods are applicable to large systems, as well as for highthroughput investigation of protein data.
2. There are some limitations in ENMs since ENMs only consider
Cα atoms without considering specific interactions. It can be
known that allostery should take into account the roles of
flexible regions such as loops. ENMs may not be suitable for
modeling the allosteric effects of these flexible regions.
3. Additionally, there are some other normal mode analysis-based
web servers to predict allosteric sites and signal propagation
pathways in proteins and their complexes, such as PARS [46],
SPACER [47], AlloPred [48], and DynOmics [49] servers.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (31872723) and a Project Funded by the Priority
Academic Program Development (PAPD) of Jiangsu Higher Education Institutions.
The author thanks Prof. Ivet Bahar for giving the opportunity
to study Elastic network models and ProDy in her lab. The author
also thanks Drs. Hongchun Li and Chakra Chennubhotla for
providing programming codes.
Identification of Allosteric Effects in Proteins by Elastic Network Models
33