2. Defining the T-Hb structure by parsing the PDB file with only
Ca atoms:
$THb = parsePDB (‘2dn2.pdb’, subset = ‘calpha’)
3. Defining the class of GNM analysis:
$gnm_T=GNM (‘T-Hb’)
4. Construction of Kirchhoff matrix of atomic coordinates:
$gnm_T.buildKirchhoff (T-Hb)
5. Calculation of GNM modes (20 modes by default) by diagonalization of Kirchhoff matrix:
$gnm_T.calcModes ()
6. Calculation of square fluctuations for the first and the second
GNM modes to identify hinge sites.
$sq_1=calcSqFlucts (gnm_T [0])
$sq_2=calcSqFlucts (gnm_T [1])
7. Saving square fluctuations for the first and the second GNM
modes.
$np.savetxt (‘sq_1.txt’, sq_1)
$np.savetxt (‘sq_2.txt’, sq_2)
The square fluctuations of the T-Hb based on the first and
second GNM modes are shown in Fig. 3a. The hinge residues for
mode 1 are distributed in the α 1 -β 2 interface, including Thr38,
Thr41, Leu91, Val93, Pro95, and Ser 138 in α 1 (the blue line),
while for mode 2 are distributed in the α 1 -β 1 interface, including
Thr118, Ala120, Ala123 in α 1 , and His112, Phe118 in β 1 (the red
line). Based on the first and second GNM modes, similar protocol
can be performed to obtain fluctuations of the R-Hb, which are
shown in Fig. 3c. The hinge residues for mode 1 (the blue line) are
not only distributed in the α 1 -β 2 interface including Val93, Asp94,
Pro95, Thr137, Ser138 in α 1 , but also the α 2 -β 1 interface including
Asp94, His97in β 1 , while for mode 2 (the red line) are also
distributed in the α 1 -β 1 interface, including Phe117, Val121 in
α1, and Ala 115 in β1. The distributions of these residues predicted
by the first GNM mode (green beads) and the second GNM mode
(yellow beads) in the three-dimensional structures of T and R-Hbs
are displayed in Fig. 3b, d, respectively.
26
Guang Hu