Generation of initial data: 1 µs simulations of the SecA-SecYEG-PP model were run with either MG.ATP or MG.ADP in the SecA nucleotide binding site – full details given in manuscript.

Analysis of SecA pore size

1. Protein coordinates were extracted from the simulations, every 25 ns from 0.75-1 µs. For our purposes, the pre-protein substrate was also removed.

2. For each snapshot, the HOLE algorithm was run (http://www.holeprogram.org).
The input script for each protein was:

coord ${NAME}.hole.pdb
radius /PATH_TO_DIR/simple.rad
cpoint $x $y $z
cvect 0.0 0.0 1.0
sphpdb hole_out_${NAME}.sph
centre info
endrad 5.

Where $NAME, $x, $y and $z are system-specific variables.

Explanation:

The analysis was initialised (cpoint) using the xyz coordinates of the centre of mass between the CA of two pore residues.
The analysis was initialized on the z axis (cvect 0 0 1)
The radius to terminate the pore was set to 0.5 nm.
A simple Amber vdw radii was used, as supplied by the HOLE program.

3. The HOLE output pore profiles were trimmed to within 0.9 nm on each side of the starting coordinates. The 10 lowest values from this region (i.e. the narrowest section of the pore) were then extracted and average to provide the smoothed pore minima

4. The smoothed pore minimum was then plotted for each snapshot, grouped by nucleotide state.