Flexibility is of great importance to a protein’s biological functioning, stability, and conformational behavior. Virtually all enzymes use some degree of flexibility to catalyze reactions, and popular docking programs include estimates of a protein’s flexibility in their calculations [1,2]. We have studied flexibility using a method outlined by Teodoro et al. , which examines the principal components of an atom’s trajectory to determine the primary axis, or direction, of its flexibility and the magnitude. Flexibility is thus related to root mean square fluctuation (RMSF) in that it measures the magnitude of motion of an atom, but is different in that it provides a directionality as well as filtering out the secondary and tertiary components of the motion, which are frequently noisy .
We performed flexibility analysis of 253 native-state protein simulations and summarized our findings in . We found that there were small but present trends in the flexibilities of secondary structure and that highly flexible regions of a protein were predictive of its early unfolding trajectory. We additionally discovered a number of loops, which were uncharacteristically inflexible despite being solvent exposed, and proposed that they be unique structural motifs.