Maintenance of electrostatic stabilization in altered tubulin lateral contacts may facilitate formation of helical filaments in foraminifera

Microtubules in foraminiferan protists (forams) can convert into helical filament structures, in which longitudinal intraprotofilament interactions between tubulin heterodimers are thought to be lost, while lateral contacts across protofilaments are still maintained. The coarse geometric features of helical filaments are known through low-resolution negative stain electron microscopy (EM). In this study, geometric restraints derived from these experimental data were used to generate an average atomic-scale helical filament model, which anticipated a modest reorientation in the lateral tubulin heterodimer interface. Restrained molecular dynamics (MD) simulations of the nearest neighbor interactions combined with a Genalized Born implicit solvent model were used to assess the lateral, longitudinal, and seam contacts in 13-3 microtubules and the reoriented lateral contacts in the helical filament model. This electrostatic analysis suggests that the change in the lateral interface in the helical filament does not greatly diminish the lateral electrostatic interaction. After longitudinal dissociation, the 13-3 seam interaction is much weaker than the reoriented lateral interface in the helical filament model, providing a plausible atomic-detail explanation for seam-to-lateral contact transition that enables the transition to a helical filament structure.


Visualizing change in lateral interaction between the 13-3 and helical filament state
To visually demonstrate the subtle lateral inter-dimer reorientation required to change from a 13-3 microtubule structure to a helical filament structure, ten short targeted molecular dynamics simulations were performed pulling the lateral 13-3 tetramer structure to the lateral helical filament tetramer structure, and vice versa. The simulation methods were the same as for the interaction energy estimations, except that the protein backbone non-hydrogen atoms were not harmonically restrained. Instead, the protein backbone atoms of only one tubulin dimer were fixed. A very strong Root Mean Square Deviation (RMSD) restraint (1000000 kcal/mol/Å 2 was imposed on all non-hydrogen atoms to change in 0.1Å increments from one state to another, with 500 steps of Steepest Descent minimization, and a 1000 steps of dynamics at each step. As can be seen in the two Supplementary Information movies, this led to motion of the second tubulin dimer relative to the fixed first tubulin dimer from the 13-3 lateral orientation to the helical filament lateral orientation, and back.

Helical filament model
Pdb format file with helical filament Cα atom model corresponding to the average cylindrical diameter and longitudinal separation in the experimental data. Can be viewed using open source molecular visualization programs (e.g. VMD).

Movie 1
Mov format file (mt133 to hf top.mov) showing the top view for five targeted molecular dynamics (TMD) forward and backward simulations between lateral contacts in a 13-3 microtubule and lateral contacts in a helical filament.

Movie 2
Mov format file (mt133 to hf side.mov) showing the side view for five targeted molecular dynamics (TMD) forward and backward simulations between lateral contacts in a 13-3 microtubule and lateral contacts in a helical filament.
x min is the minimum energy value of the restrained parameter, and x represents either the distance (b), angle (a), or dihedral (d) parameters.    Figure 2: Comparison of α-α and β-β interacting residues in 13-3 microtubule and the average helical filament model. A. Sequence and secondary structure location of residues within 5Å of neighboring dimer in the 13-3 microtubule for α-tubulin shown underlined and colored yellow and tan; B. Sequence and secondary structure location of residues within 5Å of neighboring dimer in the 13-3 microtubule for β-tubulin shown underlined and colored orange and pink; C. Residues within 5Å of neighboring dimer in a 13-3 microtubule tetramer structure; D. Sequence and secondary structure location of residues within 5Å of neighboring dimer in the helical filament for α-tubulin shown underlined and colored yellow and tan; E. Sequence and secondary structure location of residues within 5Å of neighboring protofilament in the helical filament for β-tubulin shown underlined and colored orange and pink; F. Residues within 5Å of neighboring dimer in a helical filament tetramer structure. Secondary structure elements indicated in sequences show αhelices in red, β-sheets in blue, and everything else in green. The helical filament structures are without any additional minimization, i.e. the internal structure of the monomers is the same as for the 13-3 microtubule state.     (a) A decoy foram tubulin tetramer with an incorrect inter-dimer interface obtained from the average helical filament model prior to cylindrical restraint optimization. This model complies with the coarse-grained restraints (distances or angles or dihedrals between monomer centers-of-mass), but has an incorrect orientation of the tubulin dimers with respect to the central helical axis, and with respect to each other; (B) Overall electrostatic interaction energy distributions for this interface; (C) Coulombic inter-dimer interaction energy distributions for this interface; (D) Solvation electrostatic inter-dimer interaction energy distributions for this interface. Interaction energies were obtained for a total sampling time of 1 ns for the decoy tetramer (pooled from five separate 0.2 ns simulations), and the probability distributions were generated as histograms with a bin width of either 2 kcal/mol or 10 kcal/mol.