Figure 2.

Checking the assumptions that (1) the selectivity filter remains conductive and (2) that the ions and water proceed through the selectivity filter in an alternating single file. (A) Log histogram of the backbone ψ dihedral angle of Val76 shows that the first assumption does not always hold; 2.7% of valines have a ψ > 0° that corresponds to a flipping of the backbone carbonyl oxygen and, as suggested by Bernèche,³³ leads to a selectivity filter with altered conduction properties. Because KcsA is a tetramer and only a single flip is required, 11% of frames are affected in this way. We remove these frames and those where the KwK mechanism is not followed from all subsequent analyses. (B) To check the second assumption, we have plotted the number of waters between adjacent ions as a function of the z coordinate of the ion furthest from the center of the selectivity filter. The area of each blue square is proportional to its percentage occurrence at that value of z. We further define an ion to be within the selectivity filter if it is within sites S1–S4 (gray area); for a detailed definition, see Methods. There is only a single water between any two ions in the selectivity filter for most (98%), but not all, of the time (circled in red). Waters occasionally squeeze past the potassium ions in the filter leading to unexpected configurations, two examples of which are shown.