Fig 1.
Cartoon representation of GluCl active with L-Glutamate (L-Glu) and ivermectin (IVM) bound; PDB 3RIF.
Two out of the five chains are represented in light and dark grey, respectively. The lipid membrane is materialized by grey lines so as to visualize the structural regions corresponding to the extracellular (ECD) and the transmembrane (TMD) domains. The interfacial loops β1-β2 (ECD) and M2-M3 (TMD) are shown in blue and magenta colors, respectively. The four transmembrane helices per subunit (M1 to M4) are indicated.
Fig 2.
Quaternary change of GluCl active upon removal of ivermectin (IVM).
From top to bottom, the time series of the Cα-RMSD of the TMD from the X-ray structure of GluCl apo; the receptor twisting angle; the Cα cross section of the ion pore at the position 9′; and the number of L-Glu bound are shown. All data points correspond to running averages taken over consecutive time windows of 5 ns (i.e. 500 snapshots). To ease the visualization of spontaneous pore-closing only the first 1.5 μs is shown; full-range analyses are shown in S1 Fig. Red and green dashed lines correspond to values obtained from the X-ray structures of GluCl active (PDB code 3RIF) and rest (PDB code 4TNV), respectively.
Fig 3.
The pore-closing transition of GluCl active promoted by the removal of IVM.
(A) The structural evolution of the ion pore is shown by a series of HOLE profiles computed at different time frames. Clearly, the constriction point is located at the position 9′ (Leu 254). (B) The configuration of the pore at the end of the 2.5 μs relaxation with IVM removed. The comparison with the X-ray structure of GluCl apo (green) shows a striking correlation with the MD relaxed structure in the absence of IVM (blue).
Fig 4.
Ion and water permeability in the four simulations of GluCl.
On top and middle panels, the number of water molecules and chloride ions sitting in the pore region per nanosecond is monitored over time. On bottom, the time series of the water flux through the pore in shown. Strikingly, in the simulation of GluCl with IVM removed (dark and light blue) the number of water molecules inside the pore drops from six to nearly zero in 400 ns in run A, and 800 ns in run B. In sharp contrast, when IVM is bound (red) the average number of water molecules in the pore fluctuates around six.
Fig 5.
The conformational dynamics of the EC domain.
The polar (θp) versus azimuthal (θa) components of the tilting angle of the EC subunits are plotted along the MD simulations of GluCl active (red), resting (green), and the transition from active to rest (blue). Red and green diamonds correspond to values measured in the X-ray structures of GluCl. Isocontour lines show the density of points, i.e. the higher the density, the lighter the color and are used to highlight the existence of marginally stable states sampled by MD during the relaxation with IVM removed. On top, the polar and azimuthal tilting components are illustrated using snapshots extracted from the simulations.
Fig 6.
Structural rearrangement at the EC/TM domains interface during ion-channel deactivation.
On top, the location of the strictly conserved proline (P268) at the EC/TM interface is shown. The four transmembrane helices and the position of P268 in the active (red) and resting (green) states are indicated. On bottom, the spatial distribution of the center of mass of the five P268 on the plane parallel to the membrane is shown for GluCl active with (red) and without IVM (blue), and GluCl resting (green). The center of the pore is represented by a large black dot.
Fig 7.
An upward movement of the β1-β2 loop correlated with the twisting isomerization couples orthosteric agonist unbinding to pore closing in GluCl.
On the left, the correlation between the vertical separation of the β1-β2 loop from the M2-M3 loop (ΔZ) with the twisting angle (τ) and the cross section of the pore at the constriction point are shown. ΔZ is computed as the distance, projected on the Z axis, between the α-carbons of residues P268 (on M2-M3 loop) and V45 (on β1-β2 loop) averaged over the 5 subunits. The isocontour lines correspond to the simulations of GluCl active (red) and resting (green). The color gradient from red to green illustrates the time evolution of GluCl active with IVM removed. On the right, the gating mechanism is illustrated using snapshots taken at the beginning (red) and the end (blue) of the MD relaxation. Upon L-Glu unbinding, global receptor twisting results in an upward movement of the β1-β2 loop that facilitates the passage of the bulky proline 268 at the EC/TM interface to shut the pore at position 9′.
Fig 8.
Receptor twisting versus blooming for the four simulations of GluCl.
The simulations of the end points of gating in green (resting) and red (active) sample values of the twisting and blooming reaction coordinates that are consistent with the X-ray structures of GluCl apo (green diamond) and GluCl with IVM bound (red diamond), respectively. When IVM is removed, a striking evolution of both twisting (from 15° to 22°) and blooming (from 8° to 10°) angles is observed. When only L-Glu is bound (orange) the receptor is partially twisted (τ of 17°) and contracted (θp of 8°). Isocontour lines on the relaxation of GluCl with IVM removed (blue) are used as visual guidelines to show the existence of a kinetic intermediate sampled in the early stages of the simulation. On the right-hand side, representative structures for the four simulations are shown using the same color code. The structural comparison illustrate the global character of the gating isomerization.