1.2 NSEC OF WATER MOVEMENT IN A GRAMICIDIN CHANNEL
This mpeg movie, supplements the description of water movement in the gramicidin channel as reported in Chiu, S. W., Subramaniam, S., and Jakobsson, E. 1999. Biophys. J. 76:1939-1950 The visualization was done by Elisa Ignacio, an undergraduate Bioengineering student (now just graduated) at the University of Illinois. It shows the motions over a time period of 1.2 nsec of 13 water molecules out of the 3200 in the simulated system of one gramicidin channel, 96 lipid molecules, and 3200 water molecules. These are the 13 waters that spent some time in the channel during the course of the simulation.
A good way to use this text is to leave it open in one window of your Web browser and look at the movie in another, adjacent, window. Then you can pause the movie at various points when interesting things happen that are mentioned in the text. Perhaps you will notice features in addition to the ones we have noted in the text.
At the beginning of the data collecting portion of the simulation, note that there 9 water molecules forming a hydrogen-bonded chain within the channel, one water molecule just outside the channel mouth, and three waters that are distinctly free of the channel. As references, note that 6 water molecules remained in the channel during the entire course of the simulation, 5 water molecules made a complete transition between the channel and the bulk (the three that are far from the channel at the beginning of the simulation and two that started in the channel), and two water molecules moved between the interior of the channel and just outside the channel mouth but never made a complete transition to being free of the channel.
Note that at the beginning of the simulation the channel water molecules are arranged with their dipole moments tending to point towards the bottom of the picture. A "typical" configuration has 8 or 9 water molecules in the channel forming a continuous hydrogen bonded chain with dipole moments strongly tending to orient in the same direction.
As the simulation proceeds, you will see that the water in the channel is much less translationally and rotationally mobile than water outside the channel. Also, water just outside the channel mouth is somewhat less mobile than water far away, perhaps due to interactions with the phospholipid head groups (not shown in this visualization.)
At approximately 1 second into the movie (60 psec into the simulation), one of the water molecules disappears from the bottom of the picture and reappears at the top. This is because the simulation is done with periodic boundary conditions. This "wrapping around" will occur several times during the simulation.
At approximately 3 seconds into the movie (180 psec into the simulation), notice a gap that appears in the chain of hydrogen-bonded waters, with two waters separating from the others.
Over the next 1-2 seconds of the movie (60 -120 psec of the simulation) the hydrogen-bonding pattern in the channel at the bottom of the picture is continually broken and reformed, and the pattern in the rest of the channel fluctuates substantially.
5 seconds into the movie (300 psec into the simulation) one can see a gap simultaneously form near both ends of the channel. At this point note that the dipole moments of the six water molecules forming a continuously hydrogen-bonded chain in the center of the channel have flipped--they are now pointing upward rather than down.
The full duration of the "disturbance" in the structure of the channel waters is approximately from 3 sec into the movie to approximately 6 seconds into the movie (about 180 to 360 psec into the simulation.) Prior to the disturbance, the waters in the channel form a continuously hydrogen bonded chain, tending to point downwards. Subsequent to the disturbance, the dominant water structure is the same as prior, except that the dipole moments are pointing in the opposite direction. Note however that the dipole moment projections are a statistical tendency only. At any instant the dipole moment vector of an individual water molecule may be very different from its mean value and may even point in the opposite direction of the general tendency.
Proton motion was not considered in these calculations, but an interesting corollary to these rotational fluctuations is the possibility that protons may enter the channel from either end during any time, as at any instant the water chain may present either a negative or positive face to a proton at the channel mouth.
Both before and after the disturbance, brief gaps appear occasionally in the chain of waters in the channel, and occasional episodes of hydrogen bonding with waters just outside the channel mouth also occur.