Further insights into calmodulin-myosin light chain kinase interaction from solution scattering and shape restoration.

We have gained new insight into the interactions between the second-messenger protein calmodulin (CaM) and myosin light chain kinase from skeletal muscle (skMLCK) using small-angle solution scattering and shape restoration. Specifically, we explored the nature of a 2Ca(2+)-CaM-skMLCK complex and compared it to a 4Ca(2+)-CaM-skMLCK complex under the same conditions. The 2Ca(2+) complex has been proposed to be physiologically relevant. To aid in the interpretation of the data, we developed a shape restoration approach, implemented in GA_STRUCT, that combines many of the best features of other available methods into a single, automated package. Importantly, GA_STRUCT explicitly addresses the problem of the existence of multiple solutions to the inverse scattering problem and produces a consensus envelope from a set of shapes that fit the input intensity. Small-angle scattering intensity profiles measured or calculated from known structures were used to test GA_STRUCT, which was then used to generate low-resolution models for three complexes: 2Ca(2+)-CaM-skMLCK, 4Ca(2+)-CaM-skMLCK, and 4Ca(2+)-CaM-skMLCK with a bound substrate. These models were used in conjunction with high-resolution structures of the protein components to better understand the interactions among them. In the case of the 2Ca(2+)-CaM-skMLCK complex, the consensus envelope is consistent with CaM in a fully collapsed state with its two globular lobes in close contact with each other while the catalytic cleft of the kinase is open. The consensus envelope for the 4Ca(2+)-CaM-skMLCK complex indicates that the collapsed CaM has swung further away from the open catalytic cleft of the skMLCK than in the 2Ca(2+) complex, and further that substrate binding to this complex results in closure of the kinase catalytic cleft, in agreement with previous neutron scattering results. These results indicate that activation of MLCK by CaM can only occur once CaM is fully translocated away from the catalytic cleft, which is presumably linked to full release of the pseudo-substrate/inhibitory sequence. Our scattering data indicate that this step is completed only when all four calcium binding sites are loaded.