REACTION-PATH STUDY OF LIGAND DIFFUSION IN PROTEINS - APPLICATION OF THE SELF PENALTY WALK (SPW) METHOD TO CALCULATE REACTION COORDINATES FOR THE MOTION OF CO THROUGH LEGHEMOGLOBIN

Reaction coordinates for the diffusion of carbon monoxide through leghemoglobin are calculated by using the recently developed SPW algorithm (Self Penalty Walk).1 The new algorithm makes it possible to study in a systematic way reaction coordinates in molecules with more than 1000 atoms. To explore properties of similar (but distinct) reaction coordinates, three diffusion paths were calculated. The separate coordinates were generated from different initial guesses for the paths, which were obtained from classical trajectories. 2 Analysis of the three calculated paths reveals that the ''local'' properties of the coordinates in the vicinity of the CO are very similar. The interaction energy profile of the carbon monoxide with the rest of the protein has a similar shape in the three paths, and the structural features of the local transition states are essentially the same. On the other hand, global protein properties vary considerably in the three paths. The macromolecule motions include many fluctuations that are not coupled to the diffusing ligand. It is concluded that while the density of alternative paths for the diffusion process may be very high, the close neighborhood of the ligand appears to be very much alike in the sampled paths. The diffusion process consists of two steps. In the first, the ligand hops from the heme pocket to another cavity in the protein matrix, and in the second it hops to the protein exterior. The first barrier is dominated by a tilt of a single residue (Phe 29 B9). The second barrier is more complex and includes many types of motions. In particular, global translations and rotations of helices C and G are involved.