Structure, Dynamics, and Function of S100B and S100A5 Complexes

Abstract

The S100 family is a class of small, homodimeric proteins that are often characterized by their calcium-dependent biological effects, which is typically the result of a calcium-dependent conformational change. The majority of S100 proteins have a low μM binding affinity for calcium, but in the presence of a target, this affinity can increase dramatically, as seen with the 5-fold increase in calcium binding affinity when S100B is bound to the capZ-derived TRTK-12 peptide. However, S100A5 is an exception, where the binding affinity of S100A5 for calcium is approximately 50-fold tighter than S100B not bound to a molecular target (Ca EF2KD - 0.25-1 μM). Interestingly, we have discovered that once bound to a molecular target (i.e. TXIP - Truncated eXchanger Inhibitory Peptide from NCX1) the calcium affinity for S100A5 decreases 10-fold, opposite of what is found in most other S100 proteins once bound to target. One possible explanation for the calcium "tightening" effect seen with S100B in the presence of molecular target or with S100A5 in the absence of peptide is that the calcium coordination of the EF-hands may have altered to a more optimal geometry. However, x-ray crystal structures of calcium-loaded S100B (±TRTK-12) and the calcium-bound S100A5 structure presented here, indicate that all complexes have identical calcium coordination in both the S100 (EF1) and canonical (EF2) EF-hands. Therefore a static structural explanation is not sufficient to explain how S100A5 can bind calcium so tightly in the absence of target or how calcium "tightening" occurs with S100B once bound to TRTK-12. An alternative mechanism that could explain the calcium binding properties of S100B and S100A5 may involve dynamics. For S100B, the dynamic properties for residues in the overall protein (i.e. 15N backbone amides) and EF2 (i.e. 15N side chains) could be stabilized upon S100-target complex formation. Indeed, 15N dynamics were measured for S100B in the presence and absence of TRTK-12 and upon TRTK-12 binding, the movements of several backbone amide residues were quenched at fast (ns) and slow (μs - ms) timescales. This decrease of backbone amide exchange was also translated to the EF2-hand of D63NS100B, a mutant that allows reliable detection of 15N exchange in a residue that directly coordinates calcium. For S100A5, the findings were contrary as to what was seen with S100B in the absence and bound to a molecular target. In the absence of target, there was no exchange detected in the terminal amine side chain of Asn61, a ligand that directly coordinates Ca2+ in position 3 of EF2 in S100A5. However with target bound, chemical exchange (μs - ms) and further fast time-scale motion (ns) became apparent in the backbone amides of residues in the EF-hand, helix 4 and the hinge region. These data suggest that an increase of dynamics may explain in part the decrease of Ca2+-affinity seen in the S100A5-target complex.